Renewable Energy Policy Risk and Investor Behaviour

Transcrição

Renewable Energy Policy Risk and Investor Behaviour
An Analysis of Investment Decisions
and Investment Performance
DISSERTATION
of the University of St. Gallen
Graduate School of Business Administration,
Economics, Law and Social Sciences (HSG)
to obtain the title of
Doctor Oeconomiae
submitted by
Emanuela Menichetti
from
Italy
Approved on the application of
Prof. Dr. Rolf Wüstenhagen
and
Prof. Dr. Rolf Peter Sieferle
Dissertation no. 3836
Gutenberg, Schaan
The University of St. Gallen, Graduate School of Business Administration,
Economics, Law and Social Sciences (HSG) hereby consents to the printing
of the present dissertation, without hereby expressing any opinion on the
views herein expressed.
St. Gallen, October 26, 2010
The President:
Prof. Ernst Mohr, PhD
ii
Acknowledgements
The present doctoral work has been conducted under the supervision of Prof.
Dr. Rolf Wüstenhagen, and the co-supervision of Prof. Dr. Rolf Peter Sieferle.
I would like to thank both of them for their guidance, direction and support
throughout the entire process. My deepest appreciation goes also to Prof. Dr.
Andrea Masini, HEC Paris, for his precious methodological advice, patience
and encouragement.
The work has greatly benefited from various feedbacks provided by a number
of peers, experts and scholars, whom I have met at different conferences and
seminars while carrying out my doctoral work. I thank all of them for their
valuable inputs and comments.
I also wish to thank those investors who have taken the online survey: without
their active participation, this work would have not been possible.
Furthermore, I am grateful to Mrs. Fiorella Schmucki and Mrs. Anna Schlegel,
for assisting me with the administrative aspects of my doctoral studies.
A Heartfelt thanks to my parents, Enzo and Luciana, for their unconditional
trust and support, and to my husband Paolo, whose commitment and
dedication to work make him a role model to follow for me.
Finally, the biggest thanks goes to my son Leonardo, who has given me the
right motivation and inspiration to finalise the work.
iii
Abstract
For renewable energies to achieve their market potential and contribute to
climate mitigation goals, cooperation between public and private actors needs
to be strengthened. Policy plays a paramount role in increasing investors’
confidence and decreasing uncertainty. However, the regulatory regime is
also perceived by investors as an additional source of risk. A better
understanding of the relationship between policy and market actors can lead
to improved policy design and to a more balanced estimation of risk.
Distorted risk perceptions may lead in fact to suboptimal decisions and
represent additional barriers to renewable energy market growth. Under a
microeconomic perspective, biased perceptions vis-à-vis renewable energy
technologies and policies can lead investors to overlook promising
opportunities. At a macroeconomic level, policies failing to understand the
behavioural context in which investors make decisions will not be able to
leverage enough capital in the renewable energy market.
Existing literature has emphasised the need for more quantitative studies in
this field. The present work represents one of the earliest attempts to fill this
gap. Drawing upon a multidisciplinary literature review, a conceptual model
has been developed, which investigates the role of a priori beliefs, policy
preferences and risk attitudes in influencing the decision to invest in
renewables, as well as the relationship between renewable energy share and
portfolio performance.
Based on the results of a statistical analysis using multivariate techniques with
a sample of 93 European investors, the doctoral work demonstrates that: i)
cognitive factors have a measurable influence on the decision to invest in
renewable energy technologies; and ii) there is a positive correlation between
the share of renewables and the investment portfolio performance. The study
shows which a priori beliefs mostly influence investors’ decisions, and how
investors perceive policy risk and benefits.
The main implications for scholars, policy makers and investors are discussed,
and the main recommendations for future research are drawn.
Keywords: adaptive conjoint analysis, behavioural finance, investments,
multivariate regression, renewable energy policy, risk.
iv
Zusammenfassung
Damit erneuerbare Energien ihr Marktpotential ausschöpfen und einen Beitrag
zum Erreichen der Klimaziele leisten, muss die Kooperation zwischen
öffentlichen und privaten Akteuren verstärkt werden. Die politischen
Rahmenbedingungen spielen sowohl für das Vertrauen der Investoren als
auch für den Abbau von Unsicherheiten eine herausragende Rolle. Die
Regulierung wird jedoch von den Investoren auch als eine zusätzliche Quelle
von Risiken wahrgenommen. Ein größeres Verständnis des Zusammenhangs
zwischen politischen Rahmenbedingungen und den Akteuren auf dem Markt
kann jedoch zu einer verbesserten Politikgestaltung und zu einer verringerten
Risikoerwartung beitragen.
Eine verzerrte Risikowahrnehmung kann zu suboptimalen Entscheidungen
führen, die wiederum zusätzliche Barrieren für das Wachstum des Marktes für
erneuerbare Energien darstellen. Aus mikroökonomischer Perspektive
könnten Investoren vielversprechende Investitionsmöglichkeiten von
erneuerbaren Energien auf Grund verzerrter Wahrnehmungen der
Technologien und politischen Rahmenbedingungen übersehen. Auf
makroökonomischer Ebene führen Rahmenbedingungen, die den
Verhaltenskontext, in dem Investoren Entscheidungen treffen, nicht
berücksichtigen, dazu, dass nicht genügend Kapital im Markt für erneuerbare
Energien freigesetzt wird.
Die bestehende Literatur hat die Notwendigkeit für weitere quantitative
Studien in diesem Bereich hervorgehoben. Die vorliegende Arbeit stellt einen
der ersten Versuche dar, diese Lücke zu füllen. Aufbauend auf einer
multidisziplinären Literaturrecherche wurde ein konzeptionelles Modell
entwickelt. Dieses untersucht die Rolle von a priori Annahmen, politischen
Präferenzen und Risikoverhalten bei der Entscheidung in erneuerbare
Energien zu investieren, sowie die Beziehung zwischen dem Anteil
erneuerbarer Energien und der Wertentwicklung des Portfolios.
Die Ergebnisse basieren auf einer statistischen Analyse mit multivariaten
Verfahren mit einer Stichprobe von 93 europäischen Investoren. Damit belegt
die Dissertation, dass: i) kognitive Faktoren einen messbaren Einfluss auf die
Entscheidung haben, in regenerativen Energietechnologien zu investieren,
und ii) dass eine positive Korrelation zwischen dem Anteil der erneuerbaren
Energien und der Wertentwicklung des Investitionsportfolios existiert. Die
Studie zeigt, welche vorgefaßten Überzeugungen den größten Einfluss auf
Entscheidungen der Investoren ausüben, und wie Investoren politische
Risiken und Nutzen wahrnehmen.
v
Die
Hauptauswirkungen
der
Ergebnisse
auf
Wissenschaftler,
Entscheidungsträger und Investoren werden diskutiert und die wichtigsten
Empfehlungen für zukünftige Forschung zusammengefasst.
Stichworte:
adaptive
Conjoint-Analyse,
verhaltensorientierte
Finanzierungslehre, Investitionen, multivariate Regression, Erneuerbare
Energien-Politik, Risiko.
vi
Table of Contents
Acknowledgements ................................................................................... iii
Abstract ...................................................................................................... iv
Zusammenfassung ..................................................................................... v
Table of Contents ........................................................................................ 7
List of Figures ............................................................................................. 9
List of Tables............................................................................................. 11
1
Introduction ..................................................................................... 13
1.1 Study rationale .............................................................................. 14
1.2 Methodology and research scope .................................................. 15
1.3 Expected contribution .................................................................... 16
1.4 Structure of the work ..................................................................... 17
2
The context ...................................................................................... 19
2.1 Renewable energy market trends .................................................. 19
2.1.1 Wind
.................................................................................. 19
2.1.2 Solar PV.................................................................................. 20
2.1.3 Concentrating Solar Thermal Power ........................................ 22
2.1.4 Other renewable energy technologies ...................................... 22
2.1.5 Expected evolution of the RE market ....................................... 23
2.2 Investment trends .......................................................................... 26
2.2.1 Wind
.................................................................................. 28
2.2.2 Solar
.................................................................................. 28
2.2.3 Other renewable energy technologies ...................................... 29
2.2.4 Role of different financing sources ........................................... 29
2.3 Main policies ................................................................................. 31
3
Theoretical background and literature review ................................ 34
3.1 Renewable energy investments and financial performance ............ 34
3.2 Effectiveness of policy instruments ................................................ 37
3.3 Behavioral factors affecting investment decisions .......................... 43
4
The empirical study ......................................................................... 49
4.1 Research gaps .............................................................................. 49
4.2 Research questions....................................................................... 49
4.3 Conceptual model and research hypotheses ................................. 51
5
Research design and research methods ........................................ 55
5.1 The research design...................................................................... 55
5.2 The survey instrument ................................................................... 57
5.2.1 The questionnaire structure ..................................................... 57
5.2.2 Survey administration .............................................................. 58
5.3 Quantitative research methods ...................................................... 60
5.3.1 Adaptive Conjoint Analysis ...................................................... 62
5.3.2 Multivariate regression ............................................................ 69
7
6
Descriptive statistics ....................................................................... 76
6.1 Profile of respondents.................................................................... 76
6.2 Awareness and beliefs .................................................................. 79
6.2.1 Correlations among a priori beliefs........................................... 87
6.2.2 Technology beliefs and share of investments in renewables .... 90
6.3 Main sources for informing investment decisions ........................... 94
6.4 Risk taking attitude of investors ..................................................... 97
6.5 Assessment of performance .......................................................... 99
7
Analysis of policy preferences ...................................................... 102
7.1 General findings for the sample ................................................... 102
7.2 Analysis of single policy attributes ............................................... 110
7.2.1 Level of the premium incentive .............................................. 110
7.2.2 Type of renewable energy support scheme ............................ 113
7.2.3 Duration of the support .......................................................... 117
7.2.4 Length of the administrative process ...................................... 119
7.2.5 Social acceptance ................................................................. 121
7.3 Sample segmentation .................................................................. 123
7.3.1 Influence of company type ..................................................... 123
7.3.2 Influence of the level of exposure to the RE investing domain 125
7.3.3 Influence of the share of renewables in the portfolio ............... 126
7.3.4 Influence of portfolio composition........................................... 127
7.3.5 Influence of experience ......................................................... 129
7.3.6 Influence of the background................................................... 130
7.4 Renewable energy optimism and preference for policy schemes .. 131
7.5 Degree of confidence in market efficiency and preference for policy
schemes ..................................................................................... 134
7.6 Assessing the likelihood to invest in different policy scenarios ...... 138
7.7 Sensitivity analysis: the influence of the estimate method ............ 142
8
Results of the regression model ................................................... 145
8.1 Most relevant findings for the sample ........................................... 145
8.2 Main results of the first part of the regression model .................... 152
8.3 Main results of the second part of the model ................................ 154
9
Conclusions ................................................................................... 156
10
References ..................................................................................... 164
Annex 1: The questionnaire ................................................................... 183
Annex 2: List of companies contacted................................................... 205
Annex 4: Curriculum vitae ...................................................................... 209
Statutory Declaration .............................................................................. 210
8
List of Figures
Figure 1: Shift from traditional research perspective to a bi-directional
investigation approach ................................................................................ 15
Figure 2: Wind capacity installed in the top 10 countries .............................. 20
Figure 3: Photovoltaic capacity installed in top 10 countries ......................... 21
Figure 4: Global new investment in sustainable energy 2002-2009 .............. 27
Figure 5: Conceptual model ........................................................................ 51
Figure 6: Main steps and milestones of the research design ........................ 56
Figure 7: Distribution of the list of contacts by country.................................. 59
Figure 8: Multivariate technique selection process ....................................... 61
Figure 9: Respondents’ opinions on the appropriateness of the investment
effort proposed by the ACT scenario of the ETP 2008 publication ................ 80
Figure 10: Respondents’ opinions on the appropriateness of the investment
effort proposed by the BLUE MAP scenario of the ETP 2008 publication ..... 81
Figure 11: Respondents’ opinions on the feasible share that renewables could
reach in Europe by 2050 ............................................................................. 82
Figure 12: Respondents’ opinions on the growth potential of new renewable
energy technologies .................................................................................... 83
Figure 13: Respondents’ opinions on the potential for solar energy to reach a
significant share of the world’s energy mix ................................................... 83
Figure 14: Respondents’ opinions on the large-scale growth potential of
renewable energy technologies given their intermittent availability ............... 84
Figure 15: Respondents’ opinions on the potential of PV to reach grid parity 85
Figure 16: Respondents’ opinions on the possibility for market forces to lead
to a significant exploitation of renewables .................................................... 86
Figure 17: Respondents’ opinions on the appropriateness for government to
intervene in the renewable energy sector .................................................... 86
Figure 18: Respondents’ opinions on the appropriateness of renewable
energy subsidies ......................................................................................... 87
Figure 19: Most important sources for informing investment decisions ......... 94
Figure 20: Main sources for informing investment decisions (from 1=most
important to 5=least important) .................................................................... 96
Figure 21: Distribution of the portfolio amongst PV solar technologies ......... 97
Figure 22: Forward-looking attitude of investors........................................... 99
Figure 23: Perceived past investment performance compared to direct
competitors ............................................................................................... 100
Figure 24: Expected future investment performance compared to direct
competitors ............................................................................................... 100
Figure 25: Average importance of policy attributes for the investigated sample
................................................................................................................. 103
Figure 26: Importance of the various attribute levels .................................. 104
Figure 27: Distribution of raw utilities for 100 €/MWh ................................. 110
9
Figure 28: Distribution of raw utilities for 75 €/MWh ................................... 111
Figure 31: Distribution of raw utilities for FITs ............................................ 114
Figure 32: Distribution of raw utilities for RPS, TGCs ................................. 115
Figure 33: Distribution of raw utilities for tender schemes .......................... 115
Figure 34: Distribution of raw utilities for tax incentives, investment grants . 116
Figure 35: Distribution of raw utilities for more than 20 years ..................... 117
Figure 36: Distribution of raw utilities for 10- 20 years ................................ 118
Figure 37: Distribution of raw utilities for less than 10 years ....................... 118
Figure 38: Distribution of raw utilities for less than 6 months ...................... 119
Figure 39: Distribution of raw utilities for 6-12 months ................................ 120
Figure 40: Distribution of raw utilities for more than 12 months .................. 121
Figure 41: Distribution of raw utilities for high social acceptance ................ 122
Figure 42: Distribution of raw utilities for low social acceptance ................. 122
Figure 43: Average importance of policy attributes for the sample and for the
selected sub-samples ............................................................................... 124
Figure 44: Importance of policy attributes for the sample and for RE and nonRE investors ............................................................................................. 126
Figure 45: Importance of policy attributes for the sample and for RE investors
................................................................................................................. 127
Figure 46: Importance of policy attributes for the sample and for respondents
investing only on one specific technology .................................................. 128
Figure 47: Importance of policy attributes for the sample and for the subsample split by years of experience in the RE investing domain ................. 129
Figure 48: Importance of policy attributes for the sample and for the subsample split by age of respondents ........................................................... 130
Figure 49: Importance of policy attributes for the sample and for the subsample split by academic background ....................................................... 131
Figure 50: RE optimism and confidence in RE support schemes ............... 132
Figure 51: RE optimism and confidence in feed-in tariffs............................ 133
Figure 52: RE optimism and confidence in tradable green certificates ........ 133
Figure 53: High confidence in market efficiency and preferences for TGCs 135
Figure 54: High confidence in market efficiency and preferences for FITs .. 135
Figure 55: Low confidence in market efficiency and preferences for FITs ... 136
Figure 56: Low confidence in market efficiency and preferences for TGCs . 136
Figure 57: Distribution of the sample according to the degree of confidence in
market efficiency and preferences for TGCs .............................................. 137
Figure 58: Distribution of the sample according to the degree of progovernment attitude and preferences for FITs ........................................... 138
10
List of Tables
Table 1: Synthesis of the main scenarios depicting the projected growth of
RES by 2050 .............................................................................................. 26
Table 2: Application of statistical techniques to the model variables ............. 61
Table 3: Attributes and attribute levels of renewable energy policies selected
under the present research ......................................................................... 68
Table 4: Harman’s single factor test: Eigenvalues ........................................ 72
Table 5: Harman’s single factor test: Unrotated factor patterns .................... 73
Table 6: Descriptive statistics for the sample ............................................... 76
Table 7: Significant correlations between a priori beliefs .............................. 89
Table 8: Analysis of the contingency table comparing the opinions on the
growth potential of new renewable energy technologies with the actual share
of RES in the investment portfolio ............................................................... 91
growth potential of solar energy technologies with the actual share of RES in
the investment portfolio ............................................................................... 92
large-scale growth potential of renewable energy technologies with the actual
share of RES in the investment portfolio ...................................................... 93
Table 11: Most important sources for informing investment decisions split by
age groups.................................................................................................. 95
Table 12: Most important sources for informing investment decisions split by
company type ............................................................................................. 96
Table 13: Perceived future investment performance, by age groups .......... 101
Table 14: Perceived future investment performance, by company type ...... 101
Table 15: Average importance of policy attributes for the investigated sample,
and standard deviations ............................................................................ 103
Table 16: Average utilities of policy attribute levels for the investigated sample,
and standard deviations (N=60) ................................................................ 105
Table 17: Comparison of part worth utilities between the overall sample and
the selected sub-samples.......................................................................... 124
Table 18: Comparison between the base case scenario and a scenario
characterized by a very high level of the incentive (100 €/MWh) and varying
levels of other attributes ............................................................................ 139
characterized by a high level of the incentive (75 €/MWh) and varying levels of
other attributes .......................................................................................... 141
characterized by a low level of the incentive (25 €/MWh) and a quick
administrative process .............................................................................. 142
Table 21: Average importance of policy attributes for the investigated sample,
according to the estimate method selected ................................................ 143
11
Table 22: Importance of policy attribute levels for the investigated sample, and
standard deviations calculated with HB and OLS method .......................... 144
Table 23: Descriptive statistics and Pearson Correlations .......................... 147
Table 24: Results of the multivariate regression for the first part of the
conceptual model (all coefficients are standardized) .................................. 148
Table 25: Results of the multivariate regression for the second part of the
conceptual model (all coefficients are standardized) .................................. 150
Table 26: Analysis of the investment performance: results of the logit model
................................................................................................................. 151
12
1
Introduction
In the World Energy Outlook 2009 (IEA, 2009a), the International Energy
Agency delivers an extremely clear and compelling message: current energy
policies cannot be maintained if we want to avoid severe consequences for
the climate. An “energy technology revolution” is called for, in order to meet
the challenging objective of halving CO2 emissions by 2050 compared with
2005 levels (IEA, 2008a). “The task is urgent; we must ensure that investment
decisions taken now do not leave us with inefficient, high-emitting
technologies in the long term” (IEA 2009b, page 1). It seems that a new era is
about to start, where renewable energy technologies are no longer considered
a “Cinderella option” (Grubb, 1990) but are increasingly seen as “survival
technologies” (Leggett, 2009).
Renewables have experienced a substantial growth over the last decade, both
in developed and developing countries. Nevertheless, they are far from
reaching their full potential and still account for a small fraction of the world’s
energy industry.
Fostering the transition to a low carbon society requires a significant volume
of investments in sustainable energy technologies (Meyer et al., 2009; OECD,
2008; Stern et al., 2006; UNFCCC, 2007). However, mobilizing private capital
in this field is particularly challenging in the current economic context, as
investors seem to display a certain risk aversion. As pointed out by Saponar
(2010), analysts perceive an underweight in the sector as a result of
disappointment and structural concerns.
To overcome the current challenges and help increase investors’ confidence
more tailored policies are needed. A group of 181 investment institutions
which collectively represent assets for 13 trillion US dollars has stated that
clear and appropriate long-term policy signals are essential to help investors
integrate climate change considerations into decision-making processes and
reallocate capital to low-carbon technologies (UNEP FI, 2009). Policies
therefore have a crucial role in creating more favourable conditions to
increased investments in the renewable energy market, by lowering the risk
associated to the investment and guaranteeing a stable framework. As
expressed by an anonymous financial consultant “Policies must affect
cashflow if businesses are expected to respond. Policy based on political
‘aims’ is in effect asking investors to speculate about political delivery and that
speculation, in finance terms, will demand high or venture capital level returns,
making these technologies even less attractive” (Hamilton, 2009, page 13).
13
This evidence seems to corroborate previous findings. For instance, at the
Global Ministerial Environment Forum held in Monaco in 2008 policy
representatives concluded that the main barrier to investments in greenhouse
gas mitigation technologies is not the lack of capital, but rather the lack of
appropriate policy packages to attract it (Usher, 2008a). Scaling up
investments in the renewable energy requires therefore a deeper
understanding of the relationship between policy and market actors, as more
effective policies can be designed only if the main drivers of the investment
decision making processes are thoroughly captured and assessed.
1.1
Study rationale
For renewable energies to achieve their market potential and contribute to
climate mitigation goals, cooperation between public and private actors needs
to be strengthened. Both groups play a key role. Policy makers should create
incentives to ensure that the necessary investments are undertaken (IEA,
2007). In turn, the private sector must raise substantial financial resources to
facilitate the transition towards a low-carbon economy. Policies can play a
crucial role in reducing the risk associated to an investment decision, by
providing a stable framework and decreasing market uncertainty. This is
particularly true in such a relatively young business like the renewable energy
market. The lack of familiarity in this specific domain might lead investors to
overestimate risks and overlook promising business opportunities. In this
respect, policies can correct market failures and help investors get a more
balanced perspective. However, it is also true that in policy-driven markets the
regulatory regime is often perceived by investors as an additional source of
risk. Understanding the relationship between policy and investment by looking
at investors’ attitudes in the renewable energy investing domain represents
therefore a very relevant research topic.
An increasing body of literature has started investigating how policies should
be designed to mobilize investments in the renewable energy sector. In
particular, several studies at EU level have provided a measure for policy
effectiveness. They found that some policies are better than others in
achieving the targets by ensuring lower production costs and providing a
stable and predictable framework, thus stimulating increased investment flows.
However, the majority of studies have focused only on the policy level, while
neglecting the investors’ perspective. The lack of emphasis on the investors’
side is an important shortcoming in current research, as acknowledged in
management and finance literature (Bürer and Wüstenhagen, 2008a and
2008b; Russo, 2003; Shleifer, 2000). As highlighted by the IEA (2007) the
14
response of business to policy risk is important to determine the effectiveness
of climate and energy policy, and therefore deserves further investigation.
The present doctoral work incorporates the investors’ perspective as a
relevant unit of analysis. Introducing a new angle of investigation will help to
better understand the relationship between policy and investment. This
relationship is traditionally seen as being rather unidirectional, with policy
makers having the role to set the right framework for investments to flow. As
reported in Sonntag-O’Brien and Usher (2004) “financial institutions view
themselves more as instruments of change rather than initiators”.
However, recent empirical works have provided evidence for a more complex
picture. In particular, Bürer and Wüstenhagen (2008a) found that some
investors are actively involved in policy development. The shift from the
traditional perspective to a more balanced approach looking at the interaction
between the two extremes of the relationship is reflected in the diagram below
(Figure 1).
Traditional
perspective
Policies
Investors
Alternative
model
Policies
Investors
Figure 1: Shift from traditional research perspective to a bi-directional
investigation approach
The present thesis intends to advance the knowledge on the link between
policy risk and investment behaviour. Specifically, it looks at the impact of
policy preferences, technological risk attitudes and a priori beliefs on the
decision to invest in renewable energy technologies and its implications on
investment performance.
1.2
Methodology and research scope
To help shed light on the investors’ perspective, insights from the behavioural
finance approach have been taken into account in the literature review. In fact,
cognitive and behavioural factors might hinder market’s acceptance of
renewable energy technology innovation. This can consequently limit the size
of investments in this sector and – ultimately – the success of policies.
Evidence from scholars and practitioners alike seems to suggest that
behavioural aspects deserve more attention in managerial research.
15
Based on the main research gaps identified in the literature review, the
present dissertation incorporates the behavioural finance perspective in the
analysis of the renewable energy investment decision making process, and
proposes the following research questions:
x
Do cognitive factors have a measurable influence on the decision to
invest in renewable energy technologies?
The first research question brings along a series of more specific subquestions, namely:
x
How do cognitive factors related to policy perceptions influence the
decision to invest in renewable energy technologies?, and
x
How do cognitive factors related to technology perceptions influence
the decision to invest in renewable energy technologies?
The second research question looks at the link between the share of
renewable energy technologies in the investment portfolio and portfolio
performance, and is articulated as follows:
x
How does the share of renewables resulting from these investment
decisions impact the portfolio performance?
To answer these research questions, a conceptual model has been developed,
which looks at three main categories of behavioural factors: i) a priori beliefs, ii)
policy preferences and, iii) attitudes toward technological risk.
The model has been tested using primary data collected during an online
survey with a sample of European investors and analysed through multivariate
statistical techniques. Europe was selected as an appropriate empirical
context, both for its leading role on climate change and energy policies and
because it is the world region that attracted the largest share of new
renewable energy investments in 2008 (UNEP and Bloomberg NEF, 2009).
1.3
Expected contribution
The proposed research is expected to contribute to energy policy,
management and behavioural finance literature, as well as to provide
recommendations for managerial practice. Firstly, introducing cognitive
elements in the analysis of policy effectiveness can be very enriching and lead
to a more accurate description of the relationship between policy and
investment. Thanks to a better understanding of investors’ behaviours, the
research will help policy makers to design more effective policy instruments to
support the deployment of the sustainable energy market.
16
The analysis of the influence of behavioural factors represents an important
contribution also to behavioural finance. As highlighted by Shleifer (2000) “the
perception of risk is one of the most intriguing open areas in behavioral
finance”. By looking at how and to what extent investment behaviours in the
renewable energy sector deviate from the expectations of traditional finance
theories, the research will provide new empirical insights to support this
stream of research.
Furthermore, the research is expected to advance the knowledge on the
emerging field of sustainable management research. The existing literature
has mostly focused on a restricted sample of investors, namely venture
capitalists. By expanding the scope to a broader set of investors operating in
the sustainable energy field, this work will contribute to extend previous
findings to a broader and more general context.
The present dissertation intends to contribute also to the theory of social
acceptance of renewable energy innovation. As observed by Wüstenhagen et
al. (2007), while factors influencing socio-political and community acceptance
are increasingly recognised as being important in the understanding of policy
effectiveness, market acceptance has received less attention so far. By
investigating investors’ acceptance of climate and energy policies, the present
research aims at addressing this gap.
Finally, this theme appears relevant also for practitioners, both incumbents
and new operators in the renewable energy market. Cognitive biases in
decision-making in fact create additional risk characteristics that restrain the
likelihood to raise capital funding for clean energy investments. An analysis of
cognitive biases as opposed to more rational elements of policy risks will help
financiers to get a more balanced view of policy risks and opportunities in this
promising business sector.
1.4
Structure of the work
The study is organized as follows:
x
Chapter 2 sets the context for the empirical research, by giving an
overview of the main renewable energy market, technology and
policy trends, and their expected evolution;
x
Chapter 3 reviews the most relevant literature under the framework
of the present work;
x
Chapter 4 develops the research questions and illustrates the
conceptual model elaborated to guide the empirical analysis;
17
x
Chapter 5 explains the research design and the research methods
adopted;
x
Chapter 6 reports the results of descriptive statistics for the sample
analysed;
x
Chapter 7 summarises the main results of adaptive conjoint analysis;
x
Chapter 8 summarises the main results of the multivariate regression
analysis;
x
Chapter 9 draws the conclusions and summarizes the main
recommendations for future research.
18
2
The context
The present chapter has the purpose to provide an overview of the main
trends and the expected evolution of renewable energy technologies, as
regards the following aspects: market growth, investment shares,
technological improvements. An analysis of the main policies in place in
several world regions is also provided.
2.1
Renewable energy market trends
Over the last years, renewable energy technologies have recorded
progressive growth in both developed and developing countries. According to
the WEO 2009 (IEA, 2009a) in 2007 non-hydro renewable energy sources
contributed to 3% of global electricity generation (500 TWh). Between 1990
and 2007, the share of non-hydro renewables increased from 2% to 4% in
OECD countries, and from 0% to 1% in non-OECD countries (mainly in Asia).
It is interesting to note that the same WEO in 1994 projected the contribution
of non-hydro renewables to reach 191 TWh in 2010, corresponding to 1% of
total electricity generation. Although absolute numbers are small, this
comparison suggests that renewable energy technologies have experienced
an actual increase three times higher than predicted in a relatively short time
span, also thanks to the policies implemented in an increasing number of
countries.
All technologies have witnessed substantial progresses over the last decade,
which have led also to a significant cost decrease. Growth has continued
despite the financial and economic crisis which has impacted the market over
the last two years. In the following, the main trends for single renewable
energy technologies are reported.
2.1.1
Wind
Wind power generation capacity grew by over 30% in 2009 according to the
Global Wind Energy Council (GWEC, 2010). This positive trend is triggered by
the spectacular growth recorded in the European Union, USA and China. Asia
was the world’s largest regional market for wind energy in 2009, and China
the world’s largest market.
The historical trends in wind energy installed capacity in the top 10 countries
are displayed in Figure 2.
19
Figure 2: Wind capacity installed in the top 10 countries
[Source: own elaboration based on GWEC (2009 and 2010)]
As far as the EU countries are concerned, new generation capacity in 2009
reached over 10 GW, with an annual average market growth of 23% over the
last fifteen years (EWEA, 2010). Wind installations represented 39% of total
generating capacity added in the EU in 2009, more than any other generating
technologies.
As reported by the IEA (2009b), since the early 1980s the cost of onshore
wind turbines has decreased by around a factor of three. This technology is
expected to enjoy further cost reductions thanks to technological innovation.
By 2050, onshore wind should face a total cost reduction of 23% compared to
1
current levels, with learning rates of 7% .
2.1.2
Solar PV
Although smaller than wind in absolute values, solar PV capacity is
experiencing a dramatic rise. PV is the fastest growing renewable energy
technology, with a fourteen-time capacity increase in the last 10 years, and an
annual average growth of 30% over the same period. In 2009, global
cumulative capacity reached 20 GW, therefore even surpassing the optimistic
“policy-driven scenario” target for 2010 elaborated by the European
Photovoltaic Industry Association (EPIA, 2009). Indeed, in 2009 global
1
A learning rate of 7% means that capital costs decrease by 7% for every doubling of
cumulative installed capacity.
20
capacity additions exceeded 7 GW, the largest volume of solar PV ever added
in one year (REN21, 2010).
In Europe, solar PV accounted for 16% of all new electric power capacity
additions last year. The cumulated installed capacity of solar PV in the top 10
countries is reported in Figure 3.
Figure 3: Photovoltaic capacity installed in top 10 countries
[Source: own elaboration based on EPIA (2010)]
Analysts foresee even higher growth rates in the next four to five years.
According to the latest outlook from EPIA, the world PV market could increase
up to 30,000 MW per year by 2014 (EPIA, 2010). The IEA asserts that, with
effective policies in place, solar PV could supply 5% of world electricity by
2030 and 11% by 2050 (IEA, 2010a). This means that by 2050 solar PV is
expected to equal wind in terms of contribution to electricity generation (Frankl
and Philibert, 2009).
Such promising outlook is also related to a rapid cost decrease, with learning
rates ranging between 15% and 22% (EPIA, 2009; Neji, 2007). System prices
and electricity generation costs are expected to fall by more than two-thirds by
2030 (IEA, 2010a). These estimates suggest that even solar photovoltaics,
which is currently one of the most expensive renewable energy technologies,
should achieve market competitiveness in the next 10 to 20 years thanks to a
series of technological improvements and economies of scale compared to
traditional, fossil-fuel based technologies.
21
2.1.3
Concentrating Solar Thermal Power
After a gap of more than 15 years, the Concentrating Solar Thermal Power
(CSP) market has begun to expand rapidly. Total current capacity in operation
is only around 700 MW, but there are thousands MW of CSP under
development worldwide, led by programmes in Spain and the US (IEA, 2009c).
If appropriate policies are adopted and dedicated transmission lines are
developed, CSP could become competitive for base-load electricity generation
by 2030 and produce up to 4,600 TWh per year in 2050 (IEA, 2010b).
Together, PV and CSP could represent up to 20% to 25% of global electricity
production by 2050 (Tanaka, 2010).
2.1.4
Other renewable energy technologies
Other renewable energy technologies recorded more modest growth rates (36%), although in some countries more sustained trends were observed
(REN21, 2010).
As for non waste biomass, approximately 56 GW of power capacity was in
place at the end of 2009 (REN21, 2010). Europe’s gross electricity production
from solid biomass has tripled since 2001. Leading countries are the
Scandinavian region, Austria and Germany. Significant growth in biomass
power is also seen in a number of developing countries, including Brazil, India,
Mexico, Thailand and Tanzania (REN21, 2010). According to the IEA (2009a),
cumulative capacity of biomass and waste is expected to grow by at least 5%
per year up to almost 150 GW in the coming two decades.
Geothermal capacity has increased by almost 2 GW since 2004, with an
average annual growth rate of 4%. By the end of 2009, total capacity of
geothermal power plants totaled approximately 11 GW, distributed in 24
countries. Almost 88% of that capacity is located in seven countries: the US,
the Philippines, Indonesia, Mexico, Italy, New Zealand and Iceland. It is
projected that other eleven countries – all located in Europe and the Americas
– will add geothermal plants by 2015, with global capacity reaching 18.5 GW
(REN21, 2010).
The IEA Energy Technology Perspectives (ETP) “BLUE Map Scenario”, which
aims at an energy-related CO2 emissions decrease of 50% by 2050,
envisages a world geothermal electricity generation of 1,000 TWh per year by
that date (IEA 2008a), i.e. almost 15 times higher than the current production.
Hydro remains by far the largest source of renewable electricity, with around
3,100 TWh produced in 2007 from a 920 GW capacity, representing almost
16% of total world electricity supply (IEA 2009a). An estimated 35 GW
capacity was added in 2008 and a further 31 GW during 2009, leading to
22
more than 980 GW by the end of 2009, including 95 GW of small hydro and
more than 127 GW of pumped storage capacity (REN21, 2010). The top-five
countries in terms of installed capacity are China, the US, Brazil, Canada and
Japan. Hydropower potential remains significant, but its exploitation is limited
by environmental constraints, resettlement impacts and the availability of sites.
The latter are expected to be concentrated in Brazil, China, India, Malaysia,
Russia, Turkey and Vietnam (REN21, 2010). Despite constraints, the ETP
“BLUE Map Scenario” projects a doubling of electricity production, up to
almost 5,300 TWh per year by 2050 (IEA, 2008a).
Boosted by supporting policies in OECD countries, the global demand for
liquid biofuels more than tripled between 2000 and 2007, reaching a share of
around 1.5% of total transport fuel demand (IEA, 2009d). The leader in
biofuels is Brazil, which introduced its first biofuels policy in the seventies, and
where sugarcane ethanol represents more than 50% of fuel demand for light
duty vehicles. Although having a much lower share on total domestic demand,
the US became the largest world producer of ethanol (from corn) in 2006. On
its turn, biodiesel accounted for the vast majority of biofuels consumed in
Europe. EU countries, driven by Germany, are the world leaders in biodiesel
production. Altogether, the share of biofuels on total energy demand for
transport in IEA countries increased by almost a factor six from 0.4% in 2000
to almost 2.4% in 2007 (IEA, 2009e).
These very rapid growth trends have considerably slowed down over the last
two years however, as a consequence of the emerging concerns on the
sustainability profile of biofuels. Their future deployment will depend on the
capability to rapidly develop the production of the so-called second generation
biofuels from non-edible biomass feedstocks, in particular from wooden and
agricultural residues. Considerable research efforts have been put in place in
OECD countries and major emerging economies and the technological and
economical potential is promising (IEA, 2009d).
Against these trends, global growth rates for fossil fuels have been of about 35% in recent years (REN21, 2010).
2.1.5
Expected evolution of the RE market
The crucial role of renewable energy technologies in future energy scenarios
is confirmed by the long-term projections taking into account climate mitigation
constraints. In such “alternative scenarios”, assuming at least a 50% reduction
of CO2 emissions by the half of the century, renewables are expected to
account for a significant share of total primary energy supply, to become the
23
largest source of electricity production and to be the second largest
contributor to CO2 emissions abatement after energy efficiency.
According to the IEA ETP 2008 “BLUE MAP” Scenario projections (IEA,
2008a), the share of renewables in the power sector is expected to increase
from the current 18% to almost 50% by 2050. Non-hydro technologies would
show the highest growth rate, with a twenty times increase in the next forty
years. Of these, solar is expected to grow over 400 times compared to current
levels. The same scenario projects liquid biofuels (produced exclusively from
sustainable sources) to supply 25% of total fuel demand in the transport
sector by 2050.
The scenario elaborated by Greenpeace International and EREC (2008) is
even more optimistic. Within the “energy [r]evolution scenario”, also thanks to
massive energy efficiency measures, renewables are projected to cover 60%
of primary energy demand and contribute to 77% of electricity generation by
2050. Also in this case, the growth rate of non-hydro technologies would be
the most sustained one. In particular, solar technologies would show a threethousand time increase compared to 2005 electricity production levels.
Other studies focus exclusively on the European context, and argue that a full
decarbonisation of our electricity production systems is possible with the
technologies already available, provided that a series of policy and financial
measures are implemented and that grid stability is enhanced. For example,
the roadmap issued by the European Climate Foundation (ECF, 2010) shows
the path to be followed in order to meet the G8 objective of cutting CO2
emissions by at least 80% below 1990 levels by 2050. The document clearly
highlights that achieving this target will require a profound transformation of
the European electricity generation industry which should become nearly
carbon-free.
The report points out that achieving this shift is technically feasible but will
require fundamental changes in regulation, funding mechanisms and public
support. The document also stresses that the transformation of the European
power sector would yield economic and sustainability benefits, while securing
and stabilizing Europe’s energy supply.
However, the 2050 goals will be attained only if concrete actions are taken
within the next five years. In particular, continued uncertainty about the
business case for sustained investment in low-carbon assets will impede the
mobilization of private sector capital. In order to stimulate a radical change,
the roadmap includes a series of policy imperatives for the next five years.
24
Similarly, the recent 100% Renewable Electricity Roadmap for Europe and
North Africa issued by Pricewaterhouse Coopers (PWC, 2010) underlines that
no fundamental technology breakthroughs are required to achieve such a
vision. Reaching a 100% renewable electricity goal in the two regions will
rather require a sound evolutionary development of the economic, legal, and
regulatory framework.
The roadmap depicts a 2050 vision in which the electricity supply system in
Europe and North Africa is entirely based on a renewable energy source mix
optimised per region, with mainly wind in Northern Europe, solar in North
Africa, biomass in the Baltic and Eastern Europe and hydro in the Alps and
Scandinavia. It also envisages a strongly reinforced and interconnected grid
system, aligned and cooperative energy policies and a unified European
electricity market united with the North African market.
The last EREC report “Re-thinking 2050” shows an even more aggressive
scenario envisaging a fully renewable energy mix (and not just the electricity
mix) for the European Union by 2050 (EREC, 2010). In this scenario the share
of renewable electricity in final demand increases four times from current 10%
up to 41% in 2050, mainly driven by wind and solar PV.
Furthermore, the scenario shows a massive increase of bioenergy,
geothermal and solar for renewable heating and cooling supply, with the three
sources accounting together for 45% of final energy demand in 2050. Another
10% would come from renewable transport fuels, totalling a 96% contribution
to final energy demand in 2050.
The report highlights a large set of economic, environmental and social
benefits stemming from a 100% renewable energy vision in the EU. These
include security of supply and avoided fuel costs, avoided CO 2 emissions
costs, cumulative investments and employment creation.
A synthesis of the expected trends reported in each scenario projection is
provided in Table 1.
25
Table 1: Synthesis of the main scenarios depicting the projected growth
of RES by 2050
Scenario
Scope
Target
IEA ETP 2008
Global
About 50% of RES in the
electricity generation; 25% of
biofuels in total fuel demand
in the transport sector by
2050
EREC
“Energy
[r]evolution”
Global
60% of RES in the primary
energy demand by 2050;
77% of RES in the electricity
generation
ECF
Europe
Decarbonisation
of
the
electricity industry by 2050
PWC
Europe and North
Africa
100% electricity from RES by
2050
EREC “Re-thinking
2050”
European Union
96% of RES in final energy
demand by 2050
“Blue Map”
Whatever exact shares of renewable energy will be actually reached by 2050,
all these scenarios illustrate very clearly that renewable energy technologies
have an enormous potential which is far from being fully exploited.
Nevertheless, encouraging trends have been observed over the last years in
terms of increased investment flows in the renewable energy market, as
discussed below.
2.2
Investment trends
Investments in renewable energy technologies were still negligible until the
early 2000s. As reported by UNEP and Bloomberg New Energy Finance
(2009), in 2002 new investments in sustainable energy totalled 22 USD billion,
of which non-governmental expenditures represented a minor share. Since
then, sustainable energy investments have recorded a substantial growth,
reaching almost 150 USD billion in 2007 (almost a seven-fold increase). It has
been estimated that, from 2002 until the end of 2009, the green energy market
has attracted more than USD 650 billion cumulatively (Bloomberg NEF 2009a;
UNEP and Bloomberg NEF, 2009). The years 2006 and 2007 have
26
experienced double-digit growth, whilst 2008 and 2009 have been affected by
the consequences of the global financial crisis, which has impacted the clean
energy market both directly (through a liquidity squeeze) and indirectly (via a
general fall in global energy demand associated with lower oil and gas prices).
More specifically, total investments still slightly grew in 2008 (+5%), reaching
an all-time record of 155 USD billions. After stalling in the first quarter of 2009,
investment activity in clean energy rebounded in the rest of the year and
survived the crisis, dropping only marginally to USD 145 billion at the end of
2009, as shown in Figure 4. Indeed, while some early announcements in 2009
were predicting a general fall in renewable energy investments by as much as
38% (IEA, 2009f), the consolidated results show that the drop was much more
limited – about 7% - and that a sharp recovery was experienced in the second
semester (Liebreich, 2010).
Even taking into account the slow-down in the last years, the overall period
2002-2009 showed an average annual growth rate of investments of more
than 30% per year.
Figure 4: Global new investment in sustainable energy 2002-2009
[Source: own elaboration on data from WEF 2010 and UNEP/NEF 2009]
The outlook for 2010 is positive; the figures available for the first quarter
indicate that investments were 31% higher compared to the same period of
2009, confirming the optimistic trend projected (Bloomberg NEF, 2010). The
main reasons are the impressive growth of wind farms worldwide, and the
dynamic renewable energy market development in Asia, mainly in China, that
compensated also for investment decline in Europe and North America in
27
2009. In addition, many governments have launched green fiscal measures
that have contributed to moderate the fall of investments in this industry and
increase investors’ confidence. Total budget available is estimated at around
180 USD billion, of which more than 50 USD billion is committed to
renewables. As pointed out by Liebreich (2010), only a minor share (about 9%)
of the total clean stimulus package was spent in 2009 while roughly two thirds
of the total budget should be spent during 2010 and 2011, thus contributing to
give a new impulse to renewable energy investments. Analysts forecast a
record investment in clean energy for 2010, expected to reach USD 200 billion
and possibly to continue growing beyond that level (Liebreich, 2010; WEF,
2010).
Apart from economic stimulus programmes and green fiscal measures
established by governments, this growth trend is also spurred by technology
improvements which have resulted in increased reliability and declining costs
of most renewable energy options.
Looking more in detail at the investment trends per each technology some
important differences can be observed.
2.2.1
Wind
Wind has continued to attract the highest share of new investments year after
year, thus confirming its status of mature technology. Over the five year period
2004-08 wind absorbed around 160 USD billion, i.e. around 45% of total
investment in clean energies (UNEP and Bloomberg NEF, 2009). Despite the
economic downturn, investments in wind assets grew further to more than
USD 603 billion in 2009, with a 13% increase compared to the previous year
(REN21, 2010). This corresponds to more than 60% of total global investment
in clean energy in 2009 (REN21, 2010). In the EU, for the second consecutive
year, investment in wind was larger than in any other energy technology.
Investment in wind farms in 2009 totalled €13 billion, of which €11.5 billion in
the onshore wind and €1.5 billion in the offshore wind power sector (EWEA,
2010).
2.2.2
Solar
The second most important renewable technology was solar (mainly PV),
attracting around 70 USD billion, i.e. 20% of the total over the period 20042008, with an increase of more than 50 times with respect to the investment
level in 2004 (UNEP and Bloomberg NEF, 2009). In 2009, total investments
dropped in absolute terms. This was the consequence of several factors, the
main of which being the fact that offer of PV systems became larger than
demand thanks to augmented manufacturing capacities and larger supply of
28
purified silicon feedstock. Combined with the financial crisis, difficult access to
debt finance, and the collapse of the solar PV market in Spain after a major
change in its support policy, this led to fierce competition, reduction of margins
and a very significant drop in average selling prices of PV modules and
systems (up to 50%), i.e. of specific investment costs per watt. In fact, the PV
market continued to expand in 2009 and is expected to grow further in the
coming years (EPIA, 2010).
2.2.3
Other renewable energy technologies
The biofuels sector suffered the strongest drop, with decrease in spending of
over 60% compared to the previous year and total investment returning to the
level of 2005. This was due to a combination of factors, including lower oil
prices and unfavourable market conditions for the US ethanol industry, the
credit crunch and high sugar prices in Brazil. However, the outlook for ethanol
remains positive in both countries (REN21, 2010). On the contrary, the
European biodiesel continued to stagnate and its outlook is hampered by
difficult economic conditions and new regulations, either already imposed or
expected.
As far as other renewable energy sources are concerned, while biomass and
waste experienced an increase in investments, small hydro, geothermal and
marine energy faced difficulties in raising the necessary capital to secure
projects.
2.2.4
Role of different financing sources
Private investments have become the largest source of capital for renewable
energy projects. The bulk of investments in the private sector has come from
four sectors: venture capital, private equity, public capital markets and
investment banking.
Venture capital is a major source of funding for new technology-based firms
and a decisive determinant of entrepreneurship and innovation (OECD, 2007).
Venture capital financing for renewable energy boomed during 2006/2008,
particularly for PV and biofuels, exceeding USD 3 billion worldwide in 2006
(REN21, 2008) and reaching almost USD 4.5 billion in 2008 (UNEP and
Bloomberg NEF, 2009). North America leads in venture capital investment,
with a share of around 55% of the world’s venture capital in clean energy
during the period 2006-2008.
Summing up the contribution of venture capital and private equity investors,
the total flow of investments into clean energy reached USD 7.5 billion in 2007
and its record level of 11.8 USD billion in 2008. After the third quarter of 2008
29
however, venture capital and private equity investments fell dramatically down
to their minimum in the second quarter of 2009, due to the financial crisis.
Although they recovered in the second half of the year, they totalled USD 6.6
billion, i.e. down 44% on 2008, remaining below 2007 levels (WEF, 2010).
The trend seems now upwards again, with a number of venture capital and
private equity managers closing new funds ready for investment in 2010 (WEF,
2010).
The volume of clean energy investment on the public market grew
dramatically from less than 1 USD billion in 2004 to its all-time peak of 25
USD billion in 2007. In 2008, public market new investment came back to
roughly the same level of 2006, i.e. around 13 USD billion (WEF, 2010).
In early 2009 investments came to a complete standstill, due to peak of the
financial crisis. However, the market rebounded in the rest of the year,
finishing only slightly below the 2008 investment level. Decreases in public
offerings in Europe and the US were compensated by a significant increase in
fund raising in Asia, particularly in China and Taiwan (WEF, 2010).
In the whole period 2002-2009, the financing of renewable assets accounted
for the bulk of new investment in clean energy. Asset financing ramped up
globally from 6 USD billion in 2002 to 50 USD billion in 2006, doubling again
2
up to 97 USD billion in 2008 . In that year, wind and solar accounted for
almost three quarters of total investments (respectively 44% and 28%). In total,
renewable power absorbed 83% of asset financing, biofuels 14%, with the rest
being shared between energy efficiency and other low carbon technologies
(UNEP and Bloomberg NEF, 2009).
Asset financing was hit by the financial crisis because of both lower
conventional energy prices squeezing margins, and scarcer and more
expensive capital. As for other investments, asset financing had a low-peak in
the first quarter of 2009. It rebounded in the remaining months ending at a
level of 92 USD billion at the end of the year, i.e. 5% lower than in 2008 (WEF,
2010). Again, investment decrease in Europe and Americas was offset by
increasing volumes in Asia.
Green economic stimulus measures certainly helped the recovery, although
this factor should be not overestimated, as only approximately 10% of
committed money was actually spent in 2009 (Liebreich, 2010). However, the
current level of investments remains well below the amount needed to
effectively curb CO2 emissions. Investments will need to increase up to 500
USD billion by 2030, according to WEF (2010).
2
Excluding project refinancing and small/residential projects.
30
2.3
Main policies
Specific targets and dedicated policy measures have been set by several
countries worldwide to support the deployment of renewable energy
technologies. REN21 (2010) estimates that more than 80 countries have
established renewable energy policies. There is a great variety of schemes
and implementing measures, but two main broad categories can be identified
(IEA, 2008b):
Investments support schemes, which include amongst others, capital grants
and tax exemptions; and
Operating support schemes, like for example: price subsidies, quota
obligations and green certificates, tender schemes and tax exemptions.
Operating support schemes can be further classified in price-based and
quantity-based mechanisms. Feed-in tariffs (FIT) are the most widely used
price-based support mechanism and are adopted by 50 countries and 23
States/Provinces all over the world (REN21, 2010). Under such schemes,
renewable electricity fed into the grid is rewarded with a guaranteed tariff (total
price per unit of electricity) for a determined period of time, which usually
ranges between 10 and 20 years.
More recently, Feed-in premiums (FIP) have been introduced as a more
market-oriented support instrument. In such a scheme, a premium price is
paid to the electricity producer on top of the normal electricity market price.
Both feed-in tariffs and feed-in premiums are usually technology-specific
incentives, i.e. a different tariff is paid for each specific renewable electricity
technology. In order to encourage renewable electricity technology cost
reductions and reflect economies of scale and technology learning, a
decreasing level of tariff/premium can be set.
Historically, feed-in tariff schemes have been the primary price-based policy
instrument used to support the development of renewable in Europe
(Papadopoulos et al., 1999). This type of renewable energy support scheme
was adopted by Denmark in the 1980s and by Germany and Spain in the
1990s, and is now the most popular system being adopted by 21 EU Member
3
States . The United States enacted a national feed-in law already in 1978
(REN21, 2010).
3
More specifically, three countries offer the choice between feed-in fixed prices and
premiums, and one Member State uses a pure premium incentive (EC, 2008).
31
A third typology of price-based mechanisms is represented by fiscal incentives,
such as tax exemptions or reductions. A typical example is the exemption
from carbon taxes.
As for quantity-based mechanisms, these are mainly represented by quota
obligation systems (also called Renewable Portfolio Standards) and tendering
schemes. In countries adopting quota obligation systems, producers and/or
distributors are mandated to supply a certain share of their electricity – the
target quota fixed by the government – from renewable energy sources. In
many countries, so-called tradable green certificates (TGCs) are used to
prove compliance with the obligation and to render the market more liquid:
obligated parties can either invest directly in renewable assets or buy green
certificates from other producers or suppliers. If they fail to do both, they must
pay a penalty. TGCs are generally considered more market-oriented
mechanisms, as their price is determined on the market. However, TGC
prices strongly depend on policy-makers decisions related to quota targets,
penalties, and the duration of the obligation.
Quota obligations and TGCs are technology-neutral support mechanisms and
are meant to promote the most cost-efficient renewable energy technologies
first. However, recognizing that less mature technologies also need support,
several countries have recently introduced so-called “technology banding”, i.e.
differentiated quota or incentives per technology. Quota system policies exist
at the national level in 10 countries, and have been adopted by 52 among
provinces, states and countries worldwide (REN21, 2010). In the EU, six
Member States have implemented a quota obligation system with tradable
green certificates (Ragwitz, 2010).
Finally, under tendering schemes a call for tender is issued by a national
government or other institution for the provision of a certain amount of
renewable electricity. The bidding determines competitively the price and
winning parties are usually offered standard long-term purchase contracts.
Tenders usually specify the capacity and/or production to be achieved and
can be technology- or even project/size specific (Ecofys, 2008). In the EU, this
type of policy scheme was adopted by the United Kingdom, Ireland and
France, but it was replaced by quota scheme in the United Kingdom in 2002,
and by premium feed-in tariff in Ireland in 2000. It is now used only in France
and to support offshore wind in Denmark (Weller, 2008).
There have been several changes in the renewable energy support schemes
adopted by the EU countries over the years, with feed-in tariffs being replaced
by quota systems or vice versa. For example, Belgium adopted a feed-in tariff
32
scheme for renewables until 2002, when it was replaced by a quota system
with tradable green certificates. The Netherlands introduced a quota system
with tradable green certificates in the mid 1990s, which lasted until the year
2000 and was then replaced by feed-in tariffs in 2003. In other EU countries
two or more technology-specific support schemes co-exist, as reported by
Ragwitz (2010).
33
3
Theoretical background and literature review
The previous chapter has set the context for the present study, by providing
an overview of the current status of renewables and their expected evolution,
based on the main technology, market, investment trends and the policy
support measures in place. In the present chapter, all these elements are
discussed under a theoretical perspective. Three main bodies of literature
which are relevant to the empirical work are presented: i) studies investigating
the link between investments in renewables and financial performance; ii)
studies analyzing the role of public regulators and the effectiveness degree of
policy measures, and iii) studies applying the behavioural finance perspective
to investigate the role of cognitive factors in influencing the investment
decision making process.
3.1
Renewable energy investments and financial
performance
As shown in the previous chapter, investments in renewable energy
technologies have increased steadily over the last years, and have survived
the economic downturn better than many had expected (UNEP and
Bloomberg NEF, 2010). Some experts recognise that low carbon technologies
can offer tremendous opportunities to overcome the current crisis thanks to
the numerous environmental, economic and societal benefits they incorporate
(Deutsche Bank, 2008; IEA, 2009f; Ragwitz et al., 2009). Fulton (Deutsche
Bank, 2008) identifies a “safety net effect” for clean energy investments
determined by government regulations, which ensure a built-in advantage
over most other sectors in the long term. Furthermore, the IEA called for a
“Clean Energy New Deal” to exploit the financial and economic crisis as an
opportunity to induce a permanent shift in investments to low-carbon
technologies (IEA, 2009f).
As scientific evidence on human-induced climate change becomes more
robust, and consensus over the urgency and the necessity of taking action is
reached, practitioners and policy makers are provided with a number of
dedicated tools for financial analysis and decision making. Comparative
studies on the performance of renewable energy investments versus
traditional assets are also becoming common. These studies show that, in
recent years, investments in renewable energy technologies have lead to
superior performance compared to more traditional investments (Bloomberg
NEF, 2009b; Deutsche Bank, 2009). Providing an explanation for this
phenomenon, by clarifying the underlying relations between renewable energy
34
investments and financial performance, would represent a useful theoretical
contribution and would also help policy makers design more effective policies
to attract further investments in this area. Unfortunately, management
scholars have somewhat overlooked this topic so far.
Given the lack of studies specifically addressing the link between renewable
energy investments and financial performance, some useful insights can be
obtained from studies in related fields, such as those examining the
relationship between environmental performance and financial performance.
Scholars seem to suggest that there is a positive relationship between the
companies’ involvement in environmental and social activities and their
financial performance (Cohen et al., 1995; Dowell et al., 2000; Hart and Ahuja,
1996; Porter and van der Linde, 1995; Reinhardt, 1999; Russo and Fouts,
1997). However as King and Lenox (2001) point out, correlative studies offer
only a partial picture to corporate managers and policy makers, because they
do not indicate the direction of causality. These authors therefore call for
additional research in this area, to explore how underlying firm characteristics
affect the relationship between environmental performance and financial
performance. This view is shared by Weber et al. (2005) who state that while
the positive correlation between environmental performance and financial
performance is widely accepted, the strength of the correlation and its genesis
are still often unclear.
A second shortcoming in the literature on renewable energy investments and
financial performance is that the evaluation of the profitability of various
renewable energy technology investment options is based on outdated and
inappropriate models, which create additional economic and financial
obstacles to the growth of renewables (Awerbuch, 1996; Dinica, 2006).
Some scholars argue that renewable energy technologies are already
competitive vis-à-vis their fossil alternatives, and that their value is not
correctly disclosed by levelised cost comparison techniques. This leads to a
systematic overestimation of the costs of renewable-based electricity
compared to the fossil alternatives (Awerbuch, 2003a and 2003b). Current
valuation models do not correctly reflect the real value of alternative
technologies, since they do not take into account a series of benefits of
renewable energy technologies compared to fossil fuel sources (Wright, 2002).
As a result, traditional accounting models tend to favor expense-intensive
technologies over capital-intensive technologies. As highlighted by Kaplan
(1986) accounting-based cost analysis often suggests that the incumbent
technology is to be preferred over the innovation. This might lead investors to
overlook promising technological solutions due to a lack of understanding of
35
the real value of the innovation. In the case of renewables these benefits
include reduced externalities, favourable risk profiles, flexibility in the
production and distribution process, modularity and reversibility, amongst
others. By ignoring these important attributes investors and energy planners
systematically underestimate the real value of renewable energy technologies.
In order to provide a more balanced approach, a new stream of research has
evolved which uses portfolio theory approaches (Markowitz, 1952) to evaluate
technology alternatives. Awerbuch (2000) suggests that renewable energy
technologies should not be compared based on stand-alone costs but should
be rather evaluated on the basis of portfolio cost. This is defined as a
technology’s cost contribution relative to its risk contribution to a portfolio of
generating resources. By applying this approach, the scholar finds out that a
renewable energy technology as solar photovoltaics contributes to reduce
overall portfolio risk even though it costs more on a stand-alone basis
(Awerbuch, 2000). Therefore, investment and energy planning decisions
should be taken on the basis of the evaluation of alternative resource
portfolios rather than stand-alone costs.
The application of portfolio theory approaches in the energy field has led to
insightful results. For example, Awerbuch and Berger (2003) have
investigated how renewable energy sources can contribute to risk reduction in
a mixed portfolio. The two authors suggest that adding wind, solar
photovoltaics and other fixed-cost renewables to a portfolio of conventional
generating assets leads to decreased overall portfolio cost and risk, even
though the stand-alone generating costs of renewables may be higher. For
example, in the analysis the authors report levelised annual busbar costs for
wind used of 3.97 US cents/KWh, or 1% higher than nuclear, 26% higher than
steam coal, 34% higher than oil and 44% higher than combined cycle gas
turbines. However, when looking at single cost components, it can be seen
that variable costs related to fuel and variable operation and maintenance
account for 70% in the case of gas, whereas they are zero for wind, which is
characterized by relatively high (and riskless) capital and no fuel and variable
operation and maintenance costs. The analysis suggests that if risk-adjusted
generating costs are used, renewables are most cost effective, thus resulting
in efficient portfolios with lower fossil fuel shares.
This finding has very relevant implications both for investors and energy
planners. As far as the former are concerned, renewable energy technologies
can be used in a portfolio diversification strategy in order to hedge against
fossil price fluctuations. Regarding the latter, renewables are found to be
beneficial under a macroeconomic perspective. Lind (1982) describes
36
renewable energy investments as a form of national insurance, since they
help reduce the volatility of oil and gas. Along the same lines, Awerbuch and
Sauter (2006) suggest that renewable energy technologies can help maintain
macroeconomic growth by avoiding GDP losses due to oil price volatility.
Portfolio theory has been adopted more recently for energy planning purposes.
For example, Roques et al. (2009), apply portfolio theory approaches to
assess optimal wind power deployment in Europe, by identifying optimal wind
power portfolios across a number of EU countries. Similarly, Muñoz et al.
(2009) have developed a model for minimizing investment risk and
maximizing the return of a renewable energy technology portfolio in Spain. An
overview of the most recent research applying portfolio theory to energy
planning is provided by Bazilian and Roques (2008).
All these literature contributions suggest that investment decisions should be
evaluated not only with respect to their capacity to support one specific
renewable energy technology, but, also, with respect to their ability to
guarantee a sufficiently diversified technological environment, in which radical
innovations can be developed, nurtured and, eventually, bloom. Furthermore,
it is important to understand how different factors affect renewable energy
investment decisions not only at an aggregated level, but, also, on a
technology-by-technology base because factors or policies that may favour
the deployment of one specific technology, may not necessarily be useful to
support other renewable energy options.
From this literature review, several indications for research can be derived.
Firstly, the relationship between renewable energy investments and financial
performance remains still underexamined and the strength of the correlation is
often unclear. Therefore more studies are needed in order to help clarify this
point. Secondly, aggregate models based on traditional engineering
economics do not capture the real value of renewables. This leads financiers
to underestimate renewable energy projects because of the high upfront
investments they require, despite the higher-long term potentials.
The present work intends to complement and extend this literature, by
examining how cognitive elements and behavioural factors and attitude
towards technological risk influence an actor’s willingness to invest in
renewable energy projects, by altering their perception of risk.
3.2
Effectiveness of policy instruments
Government intervention in the renewable energy market is essential to
ensure that new technologies are given the possibility to compete with
37
traditional energy sources in a more level playing field and to correct negative
externalities associated to the use of fossil fuels.
Policies can play a crucial role in reducing the risk associated to an
investment decision, by providing a stable framework and decreasing market
uncertainty. As pointed out by Ecofys (2008) “commitment, stability, reliability
and predictability are all elements that increase confidence of market actors,
reduce regulatory risks, and hence significantly reduce the cost of capital”.
Langniss (1999, pp.112) adds “in policy development, mitigating risk is
certainly an alternative to raising the level of compensation”. Indeed, risk
mitigation and risk reduction are important issues of concerns for investors;
since lowering risk leads to a reduction in the cost of capital, risk reduction
can make a larger number of projects attractive and lead to a more sustained
growth of renewable energy technologies. Therefore the role of policy makers
is to set a predictable and stable framework in order to send clear signals to
market operators.
However, the relationship between policies and investment flows is not
straightforward. Sometimes even ambitious policy targets are not able to
catalyse investments. Even worse, the policy framework is sometimes
perceived as a potential risk factor per se. In fact, policies and supporting
measures can also change, thereby affecting the profitability of investments.
As most markets for clean technologies are regulated under policy schemes,
this risk is of particular importance for renewable energy technologies (Ecofys,
2008).
An increasingly relevant body of literature is investigating the role of public
regulators, by looking at how policies should be designed to stimulate the
growth of renewable energy innovations. Early disputes mostly focused on the
effectiveness of different policy support systems in achieving policy objectives
and on their overall cost-efficiency. The literature analysis carried out does not
provide enough support to establish which support scheme is superior for
promoting the diffusion of renewable energy technologies.
For instance, several authors (Lipp, 2007, Menanteau et al., 2003, Mitchell
and Connor, 2004; Mitchell et al., 2006) have highlighted the positive role of
feed-in tariffs in leading to an increase of the renewable energy share by
lowering the risk associated with the investment decision. This positive effect
of feed-in tariff encourages the financial participation of smaller and more riskaverse investors by creating lower-risk investment conditions (Hvelplund,
2005). Other studies provide support to the hypothesis that feed-in tariffs are a
more effective policy scheme, compared to market-based approaches (Blok,
2006; Butler and Neuhoff, 2004; Contaldi et al., 2007; Couture and Gagnon,
38
2010; Madlener and Stagl, 2005; Meyer, 2003; Morthorst, 2000; Ragwitz and
Huber, 2005; Rowlands, 2005 and 2007; Sawin, 2004; Sijm, 2002). In contrast,
other analyses (e.g.: Lesser and Su, 2008) indicate that if not properly
designed, feed-in tariffs can turn to be economically inefficient since they
require policy makers to make significant guess-work as to future market
conditions and rates of technological improvement. Therefore if feed-in tariffs
are set too high or last too long, there will be inevitably a welfare loss for
society (Lesser and Su, 2008). Liebreich (2009) points out that the hidden
costs of feed-in tariffs may offset their benefits by determining liabilities in the
long term. And indeed, some experts have criticized the German feed-in tariff
scheme for being “unduly expensive” between 1990 and 1999 (Federal
Economic Ministry’s Advisory Council, 2004; in: Butler and Neuhoff, 2004).
Tradable green certificates have been introduced as an alternative to feed-in
tariffs in some EU countries since they were considered to retain the potential
to induce investments in the renewable energy sector in a more cost-efficient
way in a liberalized energy market framework (Connor, 2003; Espey, 2001;
Lauber, 2004; Morthorst, 2000 and 2001; Nielsen and Jeppesen, 2000 and
2003; Voogt et al., 2000; Wang, 2006). However, some research has
highlighted a series of significant drawbacks characterizing quota-based
models. For example, Fouquet and Johansson (2008) explain that the failure
of the British Non-Fossil Fuel Obligation (NFFO) programme was not only due
to lack of planning permissions as pointed out by the European Commission,
but also to the fact the prices were too low compared to power generation
costs. Additionally, tradable green certificate systems tend to favour the least
costly renewable energy options, thus slowing down the innovation pace.
They also increase investor risk since the income stream to be generated is
highly uncertain and may change considerably over time. This issue is of
particular importance, since the renewable energy market is mostly composed
of small and medium size investors. Similarly, Mitchell and Connor (2004)
conclude that the NFFO did not deliver deployment nor technology diversity
and was mostly beneficial only to large companies.
Bergek and Jacobsson (2010) have analysed the Swedish experience in the
implementation of a tradable green certificates system throughout the period
2003-2008. The main lessons learnt are that this market-based mechanism
has proven to be effective and cost-efficient under a societal perspective.
However, the system has not been efficient in terms of guaranteeing low
consumer costs and has also led to overcompensation to power producers,
unlike suggested by economic theory. Another drawback of the tradable green
certificate scheme in Sweden is that it has favoured the more mature
renewable energy technologies, which can better compete in a mass market.
39
As highlighted by Menanteau et al. (2003) incentive schemes for renewables
have the purpose to allow them to progress on their learning curves and to be
adopted beyond narrow niche markets. In this respect, the Sweden tradable
green certificate scheme has not met expectations regarding its capability to
stimulate technology development.
Lorenzoni (2003) made an ex ante evaluation of the Italian tradable green
certificate system, by identifying a series of unclear aspects in the policy
design which determine market uncertainty and increase the risk perceived by
investors.
An analysis of the implementation of tradable green certificate systems in
Sweden, the United Kingdom and Flanders made by Jacobsson et al. (2009)
confirms that this type of scheme tend to favour incumbent companies,
mature technologies and to generate high levels of excess profits.
The European Commission has adopted diverging attitudes as to the
capability of price-based and market-based mechanisms to ensure renewable
energy market deployment. In 1999, it issued a working paper (EC, 1999)
which advocated the implementation of quota-based systems. As reported by
several authors (Bergek and Jacobsson, 2010; Fouquet and Johansson, 2008;
Lauber, 2004), the EC preferred tradable green certificates over feed-in tariffs
as they were considered to be more competitive according to economic theory
and to be more in line with the EU single electricity market objectives. Few
years later, the EC issued a report on the support of renewable electricity (EC,
2005). This document concluded that so far feed-in tariffs have proven to be
not only effective but also the most cost-efficient renewable energy support
scheme.
To reconcile these contradictory results, scholars and researchers are
increasingly proposing hybrid combinations of the two systems, in order to
guarantee low investment risk and to attract a broad spectrum of investors
while at the same time maintaining competition, avoiding windfall profits and
stimulating technology innovation. This can be done, for example by
introducing some elements of competition in the feed-in tariff model (Lesser
and Su, 2008; Muñoz et al., 2007) through feed-in premiums – with or without
electricity price index - or by applying technology banding to green certificates
systems (Bergek and Jacobsson, 2010). Technology banding has been
introduced already in some EU countries like Italy, Romania and the United
Kingdom (Ragwitz, 2010).
As the literature review suggests, there is increasing evidence that policy
effectiveness should be measured along a wider set of attributes than the type
40
of renewable energy support scheme implemented. Indeed, as highlighted by
Haas et al. (2004, page 838): “There is no single, universally applicable ‘best’
support mechanism or policy for the bundle of different technologies known as
RES. A mix of policy instruments needs to be tailored to the particular RES
and the specific national situation to promote the evolution of the RES from
niche to mass markets. This policy mix needs to evolve with the technology”.
The authors add: “More important than the choice of the system is the proper
design and monitoring of the support system adopted; in this respect, the
functionality, stability and continuity of a policy-support system are crucial
features”. Del Río González (2008) points out that the success of the feed-in
tariff in Spain is due to the broad political commitment which has provided a
signal of stability and certainty to investors.
A study conducted by the IEA (IEA, 2008b) has found that feed-in tariffs and
tradable green certificates can be equally effective in promoting the renewable
energy market depending on the technology and on some country-specific
conditions. It is rather the adherence to key policy design principles and the
coherence of policy measures which determine the effectiveness and
efficiency of renewable energy policies. This analysis confirms previous
findings (Held et al., 2006), which have already highlighted the important role
of long term policy frameworks to enable the successful deployment of
renewable energy technologies, by providing clear signals to the market.
These analyses suggest that effective policies result from a comprehensive
design, which takes into account several attributes including, not only the type
of renewable energy support scheme in place, but also the level of the
incentive provided, the duration of the support, the administrative framework,
and social acceptance.
In this respect, Ragwitz (2010) highlights that non-economic barriers are
relevant but the quantitative relation with the effectiveness has not been yet
fully understood. This finding represents a very relevant rationale for the
current work.
The IEA identifies five key principles for effective policies aiming at fostering
the transition of renewable energy technologies towards mass market
integration (Tanaka, 2008). The two highest priorities in the short term are to
remove non-economic barriers to improve market functioning and establish
predictable and transparent support frameworks in order to attract
investments.
At the same time, it is important to ensure that incentives progressively
decrease over time in order to foster and monitor technological innovation and
41
move rapidly towards market competitiveness. Finally, in view of a massive
large-scale diffusion of renewables due consideration must be given to their
impact on overall energy systems in terms of system reliability and overall
cost efficiency.
Another important outcome of the IEA study is that specific support measures
should be developed in order to target renewable energy technologies with
different degrees of maturity. This is in line with previous recommendations by
scholars (e.g.: Christiansen, 2001).
While these studies have focused mostly on the policy level, other empirical
works have started incorporating the investor’s perspective into the picture, so
as to understand the attitude of financial actors vis-à-vis the various policy
instruments available. In particular, an investigation of the relationship
between renewable energy policy design and financing has been made by
Wiser and Pickle (1997). In their report, the authors conclude that well
designed policies can reduce renewable energy costs dramatically by
ensuring revenue certainty which, in turn, contributes to reduce financing risk
premiums. More recently, Dinica (2006) has taken an investor-oriented
perspective to analyse the diffusion potential of support systems for
renewable energy technologies. In her paper, she claims that “classifications
and analyses of support instruments’ characteristics are mainly made from the
perspective of policy makers”. She also adds that “the way financial aspects of
support systems are described is also not sufficiently suggestive with regard
to attracting potential investors”.
Bürer and Wüstenhagen (2008a, 2008b and 2009) have surveyed a sample of
60 venture capital and private equity funds to analyse investors’ preferences
for different types of support schemes. Along the same lines, Lüthi and
Wüstenhagen (2009) have examined the influence of a set of policy attributes
on the investment decisions of a sample of European PV project developers.
These empirical works provide further evidence that policy targets and the
accompanying policy instruments deployed at the national, regional and global
level have a strong influence on the investors’ decision to allocate capital to
renewable energy projects. The mentioned studies have the clear merit of
shedding further light on the relationship between policies and investment
decisions by incorporating the investors’ perspective into the picture. In this
respect, they represent an important step towards a better understanding of
the relationship between policy instruments and investment decisions. The
present work intends to complement and extend this literature, by examining
how cognitive elements and behavioural factors influence an actor’s
willingness to invest in renewable energy projects. This is an important
42
contribution, because incorporating these factors can provide a much more
accurate description of the relationship between policies and investments,
thus helping the design of better and more effective policy instruments.
3.3
Behavioral factors affecting investment decisions
Mainstream finance approaches build upon the market efficiency principle.
This concept was first introduced at the beginning of the XX century by
Bachelier but was largely overlooked until it was taken up again by the
neoclassical economist Paul Samuelson in the late 1950s. In its classic
formulation provided by Fama (1970) the efficient market hypothesis states
that in efficient financial markets prices fully reflect the available information at
any given time. The fact that rational agents share the same amount of
information at any given time makes it impossible for any investor to “beat the
market”, ergo to out-profit the others. This also implies that prices are not
predictable, but rather follow a “random walk” (Samuelson, 1965).
Three degrees of market efficiency are assumed (Fama, 1970). The weak
form hypothesis posits that no excess returns can be achieved by using
investment strategies based on historical price series. In a weak-form efficient
market current share prices are the best, unbiased, estimate of the value of
the security. The semi-weak form affirms that prices adjust to publicly
available new information very rapidly and in an unbiased fashion, such that
no excess returns can be earned by trading on that information. Finally, the
strong form states that superior returns are not possible even in the case an
investor knew inside information that is not yet available to the market. This is
because the insiders’ information quickly leaks out and is incorporated in the
prices.
Rationality of agents is at the heart of market efficiency theory, and is
embodied in the two axioms of completeness and transitivity of preferences.
Rational agents update their beliefs according to Bayes’ law.
It is worth highlighting that the theory admits occasional deviations from
rationality but even in the case some investors do not behave fully rationally,
the market as a whole is maintained in equilibrium thanks to arbitrage (Fama,
1965; Friedman, 1953).
Market efficiency and the two closely related theories of expected utility and
subjective expected utility (Savage, 1954; von Neumann and Morgerstern,
1944) have become the paradigmatic approach to describe rational behaviour
under uncertainty and have provided the foundation for prescriptive
approaches to decision making (Raiffa, 1968 and Keeney and Raiffa, 1976; in
43
Einhorn and Hogarth, 1986). They constitute the building blocks for the
development of the capital asset pricing model (Lintner, 1965; Markowitz,
1952; Mossin, 1966; Sharpe, 1964) and therefore for the theory of investor
behaviour under uncertainty.
However, over the years an increasing body of literature has been questioning
the capability of market efficiency and utility theory to provide logical answers
to a series of phenomena which provide evidence for non perfect rationality of
economic agents. One field of research which has emerged as an alternative
to the traditional market efficiency approach is “behavioural finance”.
This discipline has developed out of the pioneering work of Herbert Simon
(1957), who coined the term “bounded rationality”. The author proposed a
behavioural model of rational choice, by pointing out some cognitive
limitations of decision makers. The bounded rationality approach has been
further refined in the 1970s by Amos Tversky and the Nobel Laureate Daniel
Kahneman. The two authors applied a cognitive psychology perspective to
analyse the most common misperceptions in many decision-making
processes. Their work focused on the exploration of the systematic biases
that separate the beliefs that people have and the choices they make from the
optimal beliefs and choices assumed in rational-agent models (Kahneman,
2003). Based on a series of experiments and simulations, they found that
people tend to rely on a series of heuristics when making judgements under
uncertainty. Although these heuristics are quite useful in decision making,
sometimes they can lead to severe and systematic errors in assessing the
probability of events (Tversky and Kahneman, 1974).
In a famous work (Kahneman and Tversky, 1979), the two scholars identify a
series of behaviours that are inconsistent with the axioms of expected utility
theory. Firstly, people are found to underweight outcomes that are merely
probable in comparison with outcomes that are obtained with certainty
(certainty effect, or Allais paradox). This finding explicitly contradicts expected
utility theory’s assumption that utilities of outcomes are weighted by their
probabilities. Secondly, people are found to suffer from a reflection effect,
which leads to risk seeking attitudes when choosing between negative
prospects, and risk aversion when selecting between positive prospects. The
preferences observed in both domains contradict expected utility theory
principles as well. Thirdly, in order to simplify choices between alternatives,
people often disregard components that are shared by all alternatives under
consideration and focus on those components that differentiate them (isolation
effect). This can lead to a reversal of preferences that violates again expected
utility theory.
44
To better incorporate these deviations from full rationality, the two authors
propose a prospect theory, in which value is assigned to gains and losses
rather than to final assets and in which probabilities are replaced by decision
weights. Another insightful finding of Tversky and Kahneman’s observations is
the influence of framing on decisions which represents an additional dilemma
under a pure rational choice theory perspective (Tversky and Kahneman,
1981).
As opposed to efficient market theory, behavioural finance argues that
individuals are not fully rational (Akerlof and Yellen, 1987; Barberis and Thaler,
2003; Miller, 1977). It also argues that they do not deviate from rationality
randomly, but rather that most agents do so in similar ways. The main merit of
behavioural finance is to add a human dimension to financial market analysis.
Indeed, as highlighted by Shleifer (2000), the emphasis on investors is entirely
foreign to traditional finance.
Once considered heretical, behavioural finance has now become mainstream.
Both scholars and practitioners acknowledge the need to take into account the
limits of rationality in decision making (Lovallo and Sibony, 2010). Behavioural
finance approaches have been applied to underline a series of market
“anomalies” including the limited effects of arbitrage (Froot and Dabora, 1999;
Rosenthal and Young, 1990; Shleifer and Vishny, 1997), volatility of prices
(Shiller, 1981), overreaction and underreaction phenomena (Barberis at al.,
1998; Daniel et al., 1998; De Bondt and Thaler, 1985), the equity premium
puzzle (Barberis et al., 1999; Bernantzi and Thaler, 1995; Jegadeesh and
Titman, 1993; Mehra and Prescott, 1985), underperformance of mutual fund
managers and pension fund managers relative to passive investment
strategies (Malkiel, 1995), market’s reaction to non-information (Cutler et al.,
1991; Roll, 1984), amongst others.
The empirical evidence brought forward by these studies has challenged to
various extents the weak, semi-strong and strong efficient market hypothesis
illustrated above.
Defenders of traditional finance models reject these findings and attribute
results to operationalisational problems and non-accurate methodological
techniques. In particular, market overreaction and underreaction are plausible
phenomena within certain conditions. In fact, the efficient market hypothesis
does not exclude the possibility of short-term, occasional anomalies that might
lead an investor to outperform the market. However, these anomalies occur
randomly, so they should be attributed to chance rather than to market
inefficiency. In other terms, such phenomena are to be reconducted within the
framework of the probability law.
45
Over the years, an intense debate has been ongoing between representatives
of the two schools of thought (a good overview is offered for example in
Sewell, 2007). In particular, while Fama (1998) reiterates that market
efficiency survives the challenges from behavioural finance and states that the
alleged anomalies observed have perfectly rational explanations, Shleifer
(2000) sees these anomalies as having rather behavioural explanations.
According to Einhorn and Hogarth (1986) both utility theory and its
alternatives present some weaknesses in capturing three main elements
which characterize risky decision making. These are: i) the nature of
uncertainty in choice, ii) the effects of context, and iii) the dependence
between probabilities and payoffs. Their main criticism is that gambling as the
dominant metaphor used to conceptualize decision under risk is not a perfect
proxy for real world contexts. In addition, people are highly sensitive to
contextual variables, so that changes in context can strongly affect the
evaluation of risk. Furthermore, payoffs can affect the weight given to
uncertainty, particularly under ambiguity. This is defined as an intermediate
stage between ignorance and risk. They argue that “uncertainty about
uncertainties”, or ambiguity, is a pervasive element of much real world
decision making.
An opponent of the behavioural finance perspective is also Gigerenzer. In a
series of publications he criticizes the approach of Kahneman and Tversky by
raising objections at three levels: empirical, methodological and normative
(see the review carried out by Vranas, 2000). One of his strongest arguments
is that Kahneman and Tversky use atheoretical and rather vague terms to
label the observed heuristics, and this impedes developing a convincing
explanation on how these heuristics generate biases. He rather highlights the
importance of investigating the cognitive processes that underlie judgement
under uncertainty.
Contradictory findings are found also by behavioural finance scholars. For
example, Osborn and Jackson (1988) and Thaler and Johnson (1990) found
that past success leads to a willingness to take risks. Staw et al. (1981)
observed that when individuals are threatened by likely losses they become
more risk averse. These findings are in net contrast with the prospective
theory of Kanheman and Tversky. In their turn, March and Shapira (1987)
report a series of empirical studies which suggest that risk taking attitudes are
not connected to adversity in the simple way described by Kanheman and
Tversky.
Some scholars try to recompose the contradictions observed by combining
the complementary views of both approaches. For example, Wiseman and
46
Gomez-Mejia (1998) integrate agency theory with prospect theory to analyse
managerial risk taking attitudes. Sitkin and Pablo (1992) propose a conceptual
model that is intended to form the basis of a future research agenda regarding
the determinants of risk behaviour. Based on previous literature and additional
4
research the authors propose six determinants of risk perceptions and
examine the likely effects of risk perceptions and risk propensity on individual
risk behaviour in organisational settings.
Despite the mentioned shortcomings, behavioural finance seems to offer a
valid framework to investigate how and to what extent perceptions and biases
can influence decision making processes under uncertainty. The literature
review on behavioural finance offered above shows that motivational factors
can actually play a determinant role while weighting decision options, since
they can significantly affect the perception of risk.
As stated by Shleifer (2000, p.181) “the perception of risk is one of the most
intriguing open areas in behavioral finance”. He adds that “the emphasis on
investors is entirely foreign to traditional finance, which has achieved its
success by assuming precisely that investors do not matter except for the
determination of the equilibrium discount rate...”. The author highlights the
need to develop a conceptual model to fully capture the way investors assess
risk, their rules of thumb and how they forecast expected scenarios. In turn,
Thaler (1999) points out that adding a human element to financial market
analysis can lead to a better understanding of the mechanisms underlying
market behaviour.
The bounded rationality approach has gained recognition also in other
disciplines including economics (Van Zandt, 1999), operations management
(Bendoly, 2006; Gino and Pisano, 2008; Loch and Yaozhong, 2005), strategic
management (Bromiley, 2005), sustainability marketing (Beretti et al., 2009).
The venture capital literature (Gompers and Lerner, 2001; March, 1994;
Zacharakis and Shepherd, 2001) has acknowledged the need to better clarify
the role of cognitive factors in entrepreneurial decision making processes, as
well as these agents’ understanding of risk and return. Yet, further empirical
and theoretical work still needs to be done, particularly to study
entrepreneurial firms in the domain of sustainable technologies (Jacobsson
and Johnson, 2000; Russo, 2003).
Another recent application of behavioural finance is the analysis of the role
and effect of public policy intervention in the market. The limited available
4
The six determinants of risk perception identified by the authors are: problem framing,
top-management team homogeneity, organizational control systems, social influences,
problem domain familiarity and risk propensity.
47
literature in this new field seeks at understanding whether government
intervention, even in inefficient markets, does more harm than good.
Examples are the role of government as lender of last resort, as well as
policies to increase investor protection or to stabilise security prices.
A pan-European study in the venture capital literature published in 2003
(Leleux and Surlemont, 2003) does not provide support neither for the
seeding nor for the crowding-out effect, nor does it support the proposition
that public investments are detrimental to the industry as a whole. This implies
that further research is needed in this area, and behavioural finance can offer
a valid perspective to address these issues.
A recent paper (Allcott and Mullainathan, 2010) has highlighted the need for
more behavioural science studies in the Energy Policy field. Amongst others,
the authors recommend governments to provide funding for behavioural
programmes as part of their broader support for energy innovation. Indeed,
the analysis of the impact of behavioural factors in the energy domain has
been a largely overlooked topic. Some recent attempts to incorporate the
cognitive psychology perspective in the energy field are those of Culhane
(2008), Hasan (2006), Liang and Reiner (2009) and Piranfar (2009).
The scant literature available in renewable energy management literature offer
limited insights. For example, Teppo (2006) makes a review of some of the
most common cognitive biases identified in behavioural finance literature and
adopts a model of risky decision-making behaviour in order to examine risk
management strategies in the cleantech venture capital industry. Oschlies
(2007) incorporates the behavioural finance perspective to investigate
investment behaviours in the renewable energy market. Based on the analysis
of literature and expert interviews she develops a behavioural model of
financial decision making, which she uses to interpret the findings of a
conjoint-based experiment with a sample of institutional investors. Finally,
Wüstenhagen et al. (2009) explore the role of expectation dynamics in fuel
cell venture capital investing under a behavioural finance perspective.
While the mentioned studies represent very valid attempts to stimulate a more
thorough understanding of the influence of cognitive elements in the
renewable energy investment domain, they are still based mostly on
qualitative analysis and therefore do not help assess the nature and strength
of the relationship between behavioural factors and renewable energy
investments. To overcome this limitation and provide a contribution to this
promising field of study, I develop and test a conceptual model that examines
the behavioural factors affecting the investors’ decisions. The model is
described in detail in the next chapter.
48
4
The empirical study
The purpose of this chapter is to summarise the main knowledge gaps
identified in literature, to develop the research questions and to elaborate a
conceptual model capturing the main elements to be analysed and their
relationships.
4.1
Research gaps
The analysis of the renewable energy context, the literature review and a
series of interviews with investors and experts have allowed to identify a
series of research gaps. One important research puzzle is that although
renewable energy technologies display a very high potential for innovation
and growth (as described in chapter 2), investments in this promising field are
still below expectations.
The literature review has emphasized that bad policy design, the lack of
appropriate accounting measures for renewables and cognitive biases – such
as misperceptions, conservatorism or risk aversion - are all possible
explanations for this situation. It has also underlined that the effectiveness of a
policy is critically dependent upon its impact on investors’ behaviours.
Therefore, to maximize the impact of future energy policies, policy makers
need to get a better understanding of how investors behave, and of how they
take their decisions, particularly in regards to the key psychological factors
that may influence their actions.
Yet, there is a surprising lack of rigorous empirical studies examining these
aspects in the literature.
In the next paragraphs I try to address some of these issues, by developing
two main research questions which help define the context for the empirical
work. The research questions have the purpose to set the main boundaries of
the research and to guide the development of the research hypotheses.
4.2
Research questions
The main research questions which guide the present doctoral work are:
x
Do cognitive factors have a measurable influence on the decision to
invest in renewable energy technologies?
The first research question brings along a series of more specific subquestions, namely:
49
x
How do cognitive factors related to policy perceptions influence the
decision to invest in renewable energy technologies?, and
x
How do cognitive factors related to technology perceptions influence
the decision to invest in renewable energy technologies?
The second research question looks at the link between the share of
renewable energy technologies in the investment portfolio and portfolio
performance, and is articulated as follows:
x
How does the share of renewables resulting from these investment
decisions impact the portfolio performance?
In order to answer these research questions, a series of methodological steps
need to be undertaken.
Firstly, the most important cognitive factors under the scope of the present
research need to be identified.
Secondly, the nature and direction of the relationship between cognitive
factors and renewable energy investment have to be stated.
Thirdly, the nature and direction of the relationship between cognitive factors,
renewable energy investment and portfolio performance need to be described
as well. These are necessary steps in order to translate the problem
formulation into something testable during the field experiment.
Given the lack of a consolidated body of literature in this specific field, some
methodological indications from the grounded theory approach (Glaser and
Strauss, 1967) are followed in order to allow better interaction between theory
and research practice, create clearer links between concepts and their
indicators, and between claims and the evidence for these (Seale, 1999).
Within this context, a conceptual model has been developed, which displays
the main variables to be investigated and the relationships between them.
The main variables have been identified through primary and secondary
research, namely through stakeholders’ consultation during the preliminary
stages of the research. In order to make the link between the various
variables more explicit, a series of research hypotheses are formulated.
According to the definition provided by Kerlinger (1986, page 17) “a
hypothesis is a conjectural statement of the relation between two or more
variables”. Compared to research questions, the hypotheses allow to make
more specific predictions about the nature and direction of the relationship
between the investigated variables.
50
4.3
Conceptual model and research hypotheses
The analysis of the above literature and a series of exploratory interviews with
industry experts have provided the groundwork for the development of the
conceptual model in Figure 5.
Figure 5: Conceptual model
The model, which includes two stages, examines whether behavioural factors
have a measurable influence on the decision to invest in renewable energy
projects, and whether, in turn, the share of renewable energy in the portfolio
that results from these decisions is reflected into the portfolio performance.
The first stage examines the factors influencing an investor’s willingness to
invest in renewable energy technologies. Starting from the left-hand side of
the diagram, three main general categories of behavioural factors have been
included in the model: i) a priori-beliefs, ii) policy preferences, and iii) attitudes
towards technological risk. Therefore the model assumes that an agent’s
51
willingness to invest in renewable energy technologies is affected to various
extents by these types of cognitive elements.
As far as a priori beliefs are concerned, they are the result of the investors’
personal history, educational backgrounds, and personal previous experience
with renewable energy investments. As the success of a renewable energy
project (and the consequent appeal of such a project for an investor) depends
on the technological feasibility of the project and on the market’s ability to
value the project, the model argues that investors have a priori beliefs with
respect to both aspects. Accordingly, it is expected that investors are
influenced by two types of a priori beliefs. They reflect the investors’ trust in
the technologies considered for the investment, as well as their trust in the
efficiency of the market mechanisms (in the following defined as “market
efficiency”).
The model argues that both types of a priori beliefs have a positive impact on
the investors’ willingness to allocate capital to renewable energy projects. The
resulting hypotheses are the following:
Hypothesis 1a: The higher is the level of confidence in market efficiency the
larger is the renewable energy share in the investment portfolio.
Hypothesis 1b: The higher is the level of confidence in the technological
adequacy of renewable energy systems, the larger is the renewable energy
share in the investment portfolio.
As discussed in chapter 3, in the current energy market policies play a
paramount role in determining the success of a renewable energy project.
Thus, it is expected that an agent’s willingness to invest in a renewable
energy project will be also strongly influenced by his/her preferences over
different policy schemes. The analysis of the literature and a series of
interviews with relevant stakeholders have suggested that the effectiveness of
renewable energy policies depends on a large set of policy attributes, which
include a combination of the right policy signals, incentive levels, program
administration and predictability. By the same token, various elements of
policy play a role in determining the decision to invest in renewable energy
projects. Based on this review and the advice of industry experts, five different
policy attributes have been included in the model: i) the type of support
scheme, ii) the level of support, iii) the duration of the support, iv) the length of
the administrative process and v) the degree of social acceptance. In the
model it is assumed that the perceived importance of policy attributes strongly
influences the willingness to invest in renewable energy projects. More
specifically, the model argues that there is a positive correlation between the
52
perceived importance of the various policy attributes and the decision to invest
in renewables. The proposed hypotheses are therefore the following:
Hypothesis 2a: The renewable energy share in the investment portfolio is
positively correlated with the perceived importance of the type of policy
scheme.
Hypothesis 2b: The renewable energy share in the investment portfolio is
positively correlated with the perceived importance of the level of support.
Hypothesis 2c: The renewable energy share in the investment portfolio is
positively correlated with the perceived importance of the duration of support.
Hypothesis 2d: The renewable energy share in the investment portfolio is
positively correlated with the perceived importance of the length of the
administrative process.
Hypothesis 2e: The renewable energy share in the investment portfolio is
positively correlated with the perceived importance of social acceptance.
Finally, as renewable energy technologies are sometimes perceived as
unproven technologies with greater technological uncertainty but, also, with
the possibility to grant higher potential returns in the future, the model argues
that an investor’s attitude vis-à-vis technological risk has also a strong
influence on his/her willingness to invest. It is expected that those investors
who are more averse to technological risk are less likely to invest in
renewables compared to more risk-oriented actors. The following hypothesis
is therefore derived:
Hypothesis 3: A higher propensity to invest in radically new technologies is
associated with a higher renewable energy share in the investment portfolio.
In addition to the three main categories of behavioural factors just described,
the model includes also two control variables: the investor’s experience and
5
the type of firm undertaking the investment .
The second stage of the model considers the factors influencing the
investment performance. Controlling for the investor’s experience and the type
5
Overall, the questionnaire survey included five different parameters which could have
been used as control variables in the model: years of experience, type of company,
educational background, position in the organization and gender. All available control
variables were initially used in the model. However, for parsimony purposes, only the
ones that were significant or close to significance were eventually retained in the final
version. More specifically, gender segmentation would have not lead to significant
results, as almost all survey respondents are male. As for the other two variables, they
did not prove to be statistically relevant and were therefore eliminated from the model.
53
of firm undertaking the investment, it is expected that the performance of the
investment is dependent upon two variables: the renewable energy share in
the portfolio and the investors’ attitude towards technological risk.
As far as the first element is concerned, the literature review reported in
paragraph 3.1 has identified a positive effect of renewable energy investments
on financial performance, without however quantifying the strength of this
relationship. It has also highlighted that increased growth of renewables is
hampered by the use of traditional accounting methods which tend to favor
incumbent technologies over more innovative alternatives. Demonstrating that
a higher share of renewable energy is associated with a higher performance
of the investment portfolio would be therefore a very relevant contribution.
Based on these considerations, the following hypothesis is proposed:
Hypothesis 4: The higher is the share of renewables in the portfolio, the
higher is the investment performance.
Finally, as renewable energy technologies have a higher degree of
technological and market risks, but also the potential of guaranteeing higher
future returns, provided that these risks are properly managed, the investment
performance is expected to be positively correlated with the technological risk
attitude. The following hypothesis is derived:
Hypothesis 5: The propensity to invest in radically new technologies has a
strong impact on the investment performance.
It is worth highlighting that for sake of completeness, the existence of a
reverse feedback relationship between investment performance and the
renewable energy share in the portfolio should also be examined. In fact it can
be expected that, over time, rational investors who have obtained aboveaverage returns by including renewable in their portfolios, will also tend to
increase the share of renewable in their portfolios in the following investment
round. Therefore, it would be interesting to test the relationship between
investment performance at time t and the renewable energy share in the
investment portfolio at time t+1. Addressing this question requires an
additional experimental design which is outside the scope of the present work.
However, this limitation needs to be duly taken into account and can
constitute a valid rationale for carrying out follow-up research with longitudinal
data.
The next chapter illustrates the research design and how all the abovementioned explanatory variables are operationalised, as well as the
econometric techniques used to assess the influence of the explanatory
variables on the dependent variables.
54
5
Research design and research methods
As highlighted by several scholars (e.g.: Black, 1999; Maxim, 1999; Yin, 2003),
the research design is a fundamental step in the construction of a scientifically
sound study. In fact, a well articulated research design helps minimize
measurement errors, therefore enhancing the robustness of results (Maxim,
1999). The present chapter intends to illustrate the research design developed
under the framework of the doctoral dissertation with the aim to address the
research questions and to translate the conceptual model into empirical steps.
The main research methods adopted are explained and the main phases of
the analysis are discussed.
5.1
The research design
As defined by Yin (2003), a research design is “a logical plan for getting from
here to there, where here may be defined as the initial set of questions to be
answered and there is some set of conclusions about these questions”. In the
case of the present research, the empirical investigation is guided by the two
main questions formulated, which have in turn originated the conceptual
model and the hypotheses described in chapter 4.
Given the complexity of the problem, and the many variables to be
investigated, a research design including a combination of qualitative and
quantitative methods (Black, 1999; Snow and Thomas, 1994) has been
selected as the most suitable to best address the research questions.
As a first step, qualitative methods such as documentary analysis and
observation and hearings were adopted in order to get a solid understanding
of the main issues, priorities and problems related to renewable energy policy
and investors’ behaviours and to set the context for the empirical research.
Special attention was dedicated to the review of behavioural finance literature
in order to identify the main cognitive variables to be analysed.
The second phase of the research included interviews with selected experts in
order to test and refine the conceptual model and to assure content validity for
the various constructs in the model. A preliminary version of the web-based
questionnaire was developed and tested with a limited sample of investors
and other stakeholders. The pre-test survey helped to reformulate unclear
questions, and to refine the structure of the questionnaire by eliminating
redundant or unnecessary questions. In parallel, a database of survey
recipients was built up through extensive data collection.
55
This process allowed to finalize and launch the web-based survey
questionnaire, which was administered to a sample of European investors
during the third phase of the research. The data were then analyzed by
means of adaptive conjoint analysis and multivariate regression. Both
techniques are explained in detail below (see paragraph 5.3).
After the elaboration of the survey results, the first milestone was reached.
This consisted in presenting and discussing the findings with the survey
participants and other relevant stakeholders. In particular, a customized
complimentary copy of the study was sent to all respondents. The document
included also a feedback request module, which was used by some investors
to provide additional comments to the study results. Furthermore, the doctoral
work has been disseminated to a wide audience of scholars and practitioners,
notably through the preparation of short and full length articles for academic
conferences and peer-reviewed journals. Conferences and seminars provide
unique possibilities to present the main research findings and get very useful
feedbacks in order to improve the quality of the work.
The fourth step consisted in incorporating the various feedbacks received at
the above-mentioned events in the present dissertation, and in a series of
additional interviews with a selected number of questionnaire respondents.
Further qualitative research was carried out in order to cross-check the
findings against other empirical research in the same field. These additional
efforts led to reaching the second milestone, which is the data validation
phase, including further dissemination.
The subsequent steps of the research design are displayed in Figure 6.
Figure 6: Main steps and milestones of the research design
56
A more in-depth description of the survey process and of the quantitative
research methods used to elaborate the results is provided in the next
paragraphs.
5.2
The survey instrument
A key step of the research process was the web-based survey. To this end, a
questionnaire was developed using the Sawtooth Software SSI Web 6.6
licensed to the University of St. Gallen. The software enables the user to write
questionnaires for computer-administered interviews and surveys.
The questionnaire was sent to a list of recipients identified through extensive
data collection. In the next paragraphs, the structure of the questionnaire and
the survey administration process are described in detail. The full version of
the questionnaire is provided in Annex I.
5.2.1
The questionnaire structure
The questionnaire covered four main areas and included 35 specific questions.
The purpose of the questionnaire was both to develop valid psychometric
measurements based on Likert scales and to elicit policy preferences through
adaptive conjoint analysis. In particular, the first section aimed at determining
the diffusion of renewable energy investments. Respondents were asked to
indicate the share of their investment portfolios allocated to renewables and
the specific technologies included in their portfolios.
It is worth highlighting that a generic definition of the term “portfolio” was
adopted. In fact, this research targeted a diversified group of investors, who
have different profiles and are involved at different stages of the technology
value chain. Therefore, given the different nature of the investments these
agents undertake (e.g. stocks or bonds), it was neither possible nor
appropriate to provide a common definition for the term portfolio. In order to
avoid possible misunderstandings and to maximize the generalizability of
results, in the questionnaire respondents were explicitly invited to consider the
most representative investment undertaken by their company and to refer to
this investment.
The purpose of the second section was to assess the investors’ knowledge
and awareness of the technological and market potential of renewable energy
sources, their a-priori beliefs about the role of market and policy in supporting
the growth of renewables, as well as their attitude toward technological
innovation. A specific question addressed the main sources of information
used by respondents in order to guide their investment decision making
process.
57
As far as the question investigating market and policy a-priori beliefs is
concerned, respondents were asked to express their degree of agreement
with some statements reflecting six common beliefs about renewables, using
a 5-point Likert scale. This question was developed using a cognitive
psychology approach, employing alternative formulations of the same problem
to assess the influence of variations in framing on choice selection (Tversky
and Kahneman, 1981).
The third section was dedicated to elicit preferences for renewable energy
policies, using adaptive conjoint analysis (see paragraph 5.3.1). Respondents
were asked to compare a number of alternative policy options to support an
on-shore wind project, which differed in the type of policy scheme
implemented, level of support, duration of the support, length of the
administrative process and degree of social acceptance.
These policy attributes were selected since they fulfilled the criteria identified
by Backhaus et al. (1996). In order to reduce the possible influence of
unobserved factors and to allow for a fair comparison among attributes, some
characteristics of the project, such as the availability of wind and the project
size, were pre-defined and fully disclosed to respondents.
The fourth section of the questionnaire was dedicated to performance
assessment: respondents were solicited to provide a self-assessment of the
perceived past performance and of the expected future performance of their
investments compared to their direct competitors, i.e. other investors
operating in the same market.
Finally, the questionnaire also included a series of demographic questions
covering the company profile, the investors’ experience with renewable
energy investments, as well as their age, their educational background and
their position in the organizations.
5.2.2
Survey administration
As a first step of the data collection process, a database of target respondents
was developed in the preliminary phases of the research project. Contact
details of companies and their senior representatives were gathered from
multiple sources including the websites of the European Venture Capital
Association and its national affiliates, The Business Place website and other
specialised directories. Additional sources of information used included the
lists of participants in some of the most reputed international conferences on
sustainable energy finance, such as the Wind Energy Conference for Equity
Investors, the Renewable Energy Finance Forum, the New Energy Finance
58
Summit. Overall, a list of about 300 contacts in various European countries
was collected. The distribution of contacts is shown in Figure 7.
Figure 7: Distribution of the list of contacts by country
The database targeted mainly six European countries, which were selected
under the framework of the present research because they represent
6
particularly relevant case studies . Most contacts were found in Italy and the
United Kingdom (69 and 67, respectively) followed by France (45) while for
the remaining countries an average of more than 30 contacts per country was
reached. Fifteen additional contacts were included in the database from other
countries like for example Belgium and Greece. Although these countries
were not a primary target of the present research, selected banks and venture
capital companies were included because of their reputation and active
involvement in the renewable energy investing domain.
Investor profiles include Venture Capitalists, Private Equity Funds, Asset
Managers, Investment Funds, Commercial Banks and Energy Companies.
The administration of the survey took place between June and September
2009. Before launching the survey, a pre-test with a limited number of
investors from the sample and other relevant stakeholders was conducted in
order to validate the measurement and refine the research instrument. The
investors selected for the full scale survey received individual invitations via
email. Reminders were also sent at regular intervals. Furthermore, a link to
6
For a discussion of the country selection process, please refer to Menichetti (2008).
59
the survey was posted on the UNEP Sustainable Energy Finance Initiative
website.
In order to limit the impact of self-assessment and maximize the accuracy of
responses, Huber and Power's (1985) guidelines were followed, by
guaranteeing that the information collected would remain completely
confidential, agreeing to distribute a personalized feedback document and
promising to share the final results of the study with respondents.
The online survey was accessed 136 times. However, 43 responses had to be
7
discarded because they were either plainly unreliable or greatly incomplete.
As a result, 93 questionnaires were ultimately retained for the analysis,
corresponding to an effective return rate of 31%, which is in line with studies
of this nature. Of these, 49 questionnaires were fully completed, while 44 were
almost fully answered. To compensate for missing data, the mean substitution
method was used in order to provide all cases with complete information (Hair
et al., 1998).
5.3
Quantitative research methods
Two quantitative research methods based on multivariate data analysis were
used to elaborate the results: adaptive conjoint analysis and regression
analysis. These techniques were selected as they seem to be particularly
suited to fulfill the research objectives, according to the procedure suggested
by Hair et al. (1998) and displayed in Figure 8.
Under the present research design, each technique served a specific research
need. In particular, the conjoint analysis study had two main purposes: i) to
provide intermediate results by analyzing investors’ preferences for different
renewable energy policies, and ii) to identify the most relevant policy attributes
to be retained in the regression model. As for the multivariate regression, it
had the goal to analyse the robustness of the statistical relationships between
independent variables and the dependent variable, ergo to assess: i) the
influence of behavioural factors on the renewable energy share of the
investment portfolio and, ii) to assess the impact on the investment
performance.
In addition to these two main statistical techniques, factor analysis was
applied to investigate the interrelationships among variables and to filter a
subset of representative variables in order to simplify the subsequent
multivariate analysis and to better interpret the results.
7
For example, some problems related to common scale formats and common scale
anchors (Podsakoff et al., 2003), have been detected.
60
Figure 8: Multivariate technique selection process
[source: adapted from Hair et al. (1998, pages 20-21)]
Once identified the most suited statistical techniques for the type of
relationship to be tested, the methods shown in the diagram above have been
applied to process the variables illustrated in the conceptual model (page 51),
as summarised in Table 2.
Table 2: Application of statistical techniques to the model variables
Variables
A priori-beliefs
Policy preferences
Technological risk attitude
RE share in the portfolio
Investment performance
Assessment method
Factor analysis
Adaptive Conjoint Analysis
Factor analysis
Multivariate regression analysis
Multivariate regression analysis
61
5.3.1
5.3.1.1
Adaptive Conjoint Analysis
Introduction
Conjoint analysis is a methodological tool that allows to study consumer
preferences among multiattribute alternatives in a wide variety of product and
service contexts (Green and Srinivasan, 1978). This technique has its origin in
psychological research (Wittink and Cattin, 1989). The term “conjoint” derives
from the idea that buyers evaluate an overall product or service based on its
multiple conjoint attributes – also called features (Orme, 2009a). As opposed
to traditional expectancy-value models, conjoint methodology is characterised
by a decompositional approach (Green and Srinivasan, 1978 and 1990;
Jaeger et al., 2000). In other terms, products or services are thought as
possessing specific levels of defined attributes, and respondents’ liking for a
product is modelled as the sum of the part worths (utilities) for each of its
attribute levels. Therefore the purpose of the research is to determine the
contributed portion (part-worth utility) of each attribute level to the dependent
variable (Moore, 1980).
Conjoint analysis assumes that a consumer assigns a utility value to each
level of each attribute and makes his or her final decision based on the total
utility values across attributes for a given choice set (Green and Srinivasan,
1978 and 1990; Jaeger et al., 2000; Marshall and Bradlow, 2002; Randolph
and Ndung’u, 2000). Assuming that a product can be defined as a vector in a
multidimensional attribute space, and that the evaluation of the product is
based on its attribute levels, it becomes theoretically possible to relate
preference to attributes (Janssen et al., 1991).
In contrast to direct questioning methods that simply ask how important each
attribute is or the desirability of each level, conjoint analysis forces
respondents to make tradeoffs like the ones they encounter in the real world.
It would be time consuming and difficult for respondents to evaluate all
possible product combinations in order to provide information on their values
for the various product features. Conjoint analysis offers the researcher a
more efficient way to obtain such information: only a careful chosen set of
hypothetical product concepts is presented to respondents for evaluation. It is
the job of the analysts to find a set of parth worths for the individual attributes
that, given some type of composition rule, are most consistent with the
respondent’s overall preferences. By analysing the answers conjoint analysis
can estimate the weights and preferences respondents may have placed on
the various features in order to result in the observed product preferences.
62
Conjoint analysis has been extensively used in applied psychology, as well as
in marketing research. Over the years, the methodology has acquired
popularity also in other academic disciplines and among practitioners, as
described in the next paragraph.
5.3.1.2
Origin and evolution of the methodology
Conjoint measurement traces its root back to 1920s, but the 1964 seminal
paper of Luce and Tukey appeared in the first issue of the Journal of
Mathematical Psychology (Luce and Tukey, 1964) is generally acknowledged
to be the cornerstone of the methodology development. Following this
contribution, a number of theoretical works were released (Krantz, 1964;
Tversky, 1967), dealing either with algorithmic developments (Carroll, 1969;
Kruskal, 1965; Young, 1969) or methodological applications (for a
comprehensive review see Green and Srinivasan, 1978 and 1990). There are
several methodological papers available, which provide a comprehensive
description of the methodology and the mathematical algorithms behind the
tool (see for example: Green and Rao, 1969; Green and Wind, 1973 and 1975;
Rao, 1977).
Conjoint analysis has been extensively used in applied psychology, as well as
in marketing research where it was first introduced by Green and Rao (1971)
and further developed by Batsell and Lodish (1981) and Louviere and
Woodworth (1983). Over the years, the methodology has acquired popularity
not just in the academia but also among practitioners and has extended its
application boundaries to several other fields thanks to its possibility to
simulate the decision making process in real life situations.
Jaeger et al. (2000) report examples of application in several sectors,
including transportation, tourism and recreation, environmental valuation and
shopping behaviour. In the sustainability marketing and consumer behaviour
domain, conjoint analysis has been applied quite extensively. For example it
has been used to study the importance of green packaging (Rokka and
Uusitalo, 2008), to evaluate the influence of ecolabelling on purchase
decisions (Heinzle and Wüstenhagen, 2009; Sammer, 2007) or to determine
household preferences and willingness to pay for energy-saving measures
(Banfi et al., 2008; Farsi, 2010; Poortinga et al., 2003).
A review of the literature on the application of conjoint analysis in the
environmental field made by Alriksson and Öberg (2008) identified a total
number of 84 studies related to agriculture, ecosystem management, energy,
environmental evaluation, forestry, land management, pollution, products,
recreation, environmental risk analysis and waste management. The authors
63
conclude that compared to more traditional areas of application like marketing
and transportation, the number of environmental conjoint studies is rather
small but increasing, and that the method seems to work effectively in eliciting
preferences on environmental issues.
Conjoint analysis has been widely applied also to measure public acceptance
of policies. For example, since the early 1970’s, it has been applied for health
care planning (McClain and Rao, 1974; Whitmore and Cavadias, 1974; Wind
and Spitz, 1976; Hopkins et al., 1977).
In the energy policy context, it has been used to assess public preferences
over nuclear power (Sung Choi and Whi Lee, 1995) or wind (Álvarez-Farizo
and Hanley, 2002).
In strategic management literature, this technique has been used to analyse
investment decision making processes of entrepreneurs like venture
capitalists (Franke et al., 2006; Muzyka et al., 1996; Riquelme and Rickards,
1992; Shepherd and Zacharakis, 1999), informal investors (Landström, 1998),
management buyout investors (Birley et al., 1999), to help design rural credit
market (Dufhues et al., 2004) as well as to investigate strategic thinking of top
managers (Hitt and Tyler, 1991; Priem, 1992; Priem and Harrison, 1994).
To the author’s knowledge the present dissertation represents one of the
earliest attempts to apply this technique to assess investors’ preferences over
renewable energy policy characteristics (a similar experiment was conducted
by Lüthi and Wüstenhagen, 2009).
5.3.1.3
Brief description of main steps
A conjoint analysis study includes the following key steps:
x
Attribute List Formulation
A business problem is defined and an attribute (features) list is developed to
study the problem.
x
Data collection
Respondents are asked to express the trade-offs they are willing to make
among product features by rating, sorting or choosing among hypothetical
product concepts.
x
Utility calculation
A set of preference values (also called part worth utilities) is derived from the
interview data; they reflect the trade-offs each respondent made.
64
x
Market Simulation
The utility values are used to predict how buyers will choose among
competing products and how their choices are expected to change as product
features and/or price are varied.
8
Among the various conjoint measurement techniques available Adaptive
Conjoint Analysis - ACA (Johnson, 1987) was selected as the most suitable
approach under the scope of the present study since it retains the following
main features: a) it collects preference data in a computer-interactive mode,
therefore increasing the respondent’s interest and involvement with the task,
and b) it allows to customize the interview so that respondents are asked in
detail only about those attributes of greatest relevance.
The term adaptive refers to the fact that the computer-administered interview
is customized for each respondent. At each step previous answers are used
to decide which question to ask next, to obtain the most information about the
respondent’s preferences (Sawtooth, 2007). The respondent’s part worths are
continually re-estimated as the interview progresses, and each question is
chosen to provide the most additional information, given what is already
known about the respondent’s values. Compared to the other conjoint
methods described, ACA has the advantage to provide statistically significant
results also with a limited sample and to allow for the inclusion of a relatively
high number of attributes and levels.
The ACA interview is usually composed of several sections, each one having
a specific purpose. As described in Orme (2009a) and Sawtooth (2007), the
main steps are the following:
x
Preference for Levels (the “ACA Rating” question type)
As a first step, respondents are asked to rate the levels in terms of relative
preference. This question is usually omitted for attributes for which the
respondents’ preferences are obvious, like for example price or quality. In
such cases, the researcher can set a predefined order of preference for levels,
from “best to worst” or “worst to best”, and the rating question is automatically
skipped.
8
There are several conjoint measurement techniques, ranging from traditional fullprofile methods to two-factor methods to hybrid methods (Gustafsson et al., 2007;
Huber et al; 2007). In the present research, a review of the most popular has been
carried out, which included conjoint value analysis (CVA), choice-based conjoint
analysis (CBC), adaptive conjoint analysis (ACA). For a comprehensive description of
the various methodological tools see the Sawtooth Technical Papers library, available at:
http://www.sawtoothsoftware.com/education/techpap.shtml, and in particular the paper
of Orme (2009b).
65
x
Attribute Importance (the “ACA Importance” question type)
This step has the purpose to learn the relative importance of each attribute to
respondents. This provides information upon which to base initial estimates of
respondents’ utilities. The questioning is based on differences between those
levels the respondent would like the most and the least.
Similarly to the rating question, also the importance question can be omitted.
However, an ample size is needed in this case, as some information at the
individual level is lost (King et al., 2004).
x
Paired-Comparison Trade-Off questions (the “ACA Pairs” question
type)
In this phase of the analysis, a series of customized paired-comparison tradeoff questions is presented. The respondent is shown two product concepts at
a time, and is solicited to indicate which one is preferred, and the strength of
preference.
The computer starts with a crude set of estimates for the respondent’s utilities,
and updates them following each pairs question. The crude estimates are
constructed from the respondent’s preference ranking or rating for levels, and
the importance ratings of attributes. Every time the respondent completes a
pair question, the estimate of the respondent’s utility is updated, thus
improving the quality of the subsequent pairs questions.
x
Calibrating Concepts (the “ACA Calibration” question type)
As a last step, a series of calibrating concepts are presented, using those
attributes determined to be most important. These concepts are chosen to
occupy the entire range from very unattractive to very attractive for the
respondent. The respondent is asked to indicate the likelihood of buying each
of them.
The purpose of this section is to scale the utilities non-arbitrarily, so that sums
of utilities for these concepts are approximately equal to logit transforms of the
respondent’s likelihood percentage.
Calibrating concepts are not required in case of share of preference
simulations, and in particular if using the ACA/HB system for hierarchical
Bayes estimation, as in the case of the present research.
66
5.3.1.4
Application to the present study
The conjoint analysis design was elaborated following the procedure
suggested by Backhaus et al. (1996) and already applied by Oschlies (2007).
Conjoint analysis is based on the assumption that each “product” is composed
of an almost infinite number of attributes. However, not all attributes are
relevant to different customers. Therefore an important step in conjoint
measurement is to define which attributes and levels do influence the most
customers’ preferences.
This screening process requires a participatory process with the respondents
to be carried out in the preliminary stages of the research. To do so, an
orthogonal matrix has been developed, which allowed to screen and reduce
the number of attributes and levels in order to keep the number of product
concepts to a manageable size.
According to the approach defined by Backaus et al. (1996), attributes shall
fulfill the following criteria:
x
Relevance (they need to be relevant in the investor’s decision
making)
x
Influence (they can be influenced by investors)
x
Independence (attributes and levels should not interact)
x
Compensatory (can be substituted one with the other in investor’s
perceptions)
x
Understandable (not misleading)
By applying the above-mentioned criteria to the list of attributes identified in
the preliminary stages of the research process (Menichetti, 2008), only five
attributes were ultimately retained. They were fine-tuned after the pre-test
survey, which helped better specify attributes and levels in order to avoid
misinterpretation.
The final list of attributes and attribute levels is displayed in Table 3.
The combination of attributes and levels led to a total number of 14 paired9
comparison trade-off choices to be answered by investors .
9
The number of trade-off questions is determined according to the following formula:
3*(N-n-1)-N where N=total number of levels taken into the pairs section, n=total number
of attributes taken into the pairs section (Orme 2009a, p.324).
67
Table 3: Attributes and attribute levels of renewable energy policies
selected under the present research
Attributes
Levels
1. Tax incentives / Investment grants
1. Type of renewable
energy support scheme
2. Tender schemes
3. Feed-in tariffs
4. Tradable Green Certificates / Renewable
Portfolio Standard
2. Level of the incentive
(the premium paid per
unit of electricity
produced and sold)
1. 100 €/MWh incentive
3. Duration of the support
(number of years for
which the incentive is
paid)
1. incentive unchanged for less than 10 years
4. Length of the
administrative process
2. incentive unchanged from 10 to 20 years
3. incentive unchanged for more than 20
years
1. Less than 6 months
2. From 6 to 12 months
3. More than 12 months
1. Low (anti-wind activism, negative press,
anti-wind demonstrations)
5. Social acceptance
2. High (pro-wind activism of NGOs,
favourable press, pro-wind citizens’
coalitions)
The survey proceeded as follows: first, a series of “importance questions”
were asked by the software, where the highest and the lowest levels of each
attribute were compared. Then, a series of “pairs questions” were
automatically generated by the software, where trade-off investment decisions
were proposed. Two attributes at a time were shown. Although concepts
described on several attributes have the advantage of seeming more realistic,
the feedbacks received during the pre-test phase indicated that respondents
were tired and tended to be confused if presented with more than two
attributes, as they had to process more information. Preliminary evidence in
conjoint research suggests indeed that beyond three attributes gains in
efficiency are offset by respondent confusion due to task difficulty (Sawtooth,
2007). In order to avoid non-accurate responses, it was then decided to
present only two attributes at a time.
68
Finally, following the methodological suggestions provided by Sawtooth (2007)
and already presented in paragraph 5.3.1.3, both rating and calibration
questions were omitted. In fact, since respondents’ preferences for attribute
10
levels were rather obvious , pre-defined orders of preferences for levels were
set. In addition, since the hierarchical Bayes estimation method was used to
process the data it was not necessary to add the calibration questions.
As already anticipated, the conjoint analysis study had two main purposes: i)
to provide intermediate results by analyzing investors’ preferences for different
renewable energy policies, and ii) to identify the most relevant policy attributes
to be retained in the regression model, as discussed in the next paragraph.
5.3.2
5.3.2.1
Multivariate regression
Introduction
Regression analysis is by far the most widely used and versatile dependence
technique, applicable in every facet of business decision making (Hair et al.,
1998).
Multiple regression analysis is a general statistical technique used to analyse
the relationship between a single dependent (criterion) variable and several
independent (predictor) variables.
The objective of regression analysis is to predict a single dependent variable
from the knowledge of one or more independent variables.
Its basic formulation is:
Y1
X 1 X 2 ... X n
In principle, this statistical technique should be used only when both the
dependent and independent variables are metric. However under certain
circumstances nonmetric data can be included as well, either as independent
variables (as dichotomous variables, known as dummy variables) or the
dependent variable. As explained more in detail below, three nominal
variables were included in the model through dummy-variable coding (see
next paragraph).
As other multivariate techniques, regression analysis is a very valuable tool to
conduct theoretically significant research, and to evaluate the effects of
naturally occurring parametric variations in the context in which they normally
occur (Hardyck and Petrinovich, 1976). Thanks to its flexibility, it can be used
both for predictive and explanatory purposes. It allows to objectively
10
For example, it can be expected that a premium incentive level of 100 €/MWh is
always preferred to a level of € 75/MWh or lower.
69
determine the degree and character of the relationship between the
dependent variable and the independent variables. Firstly, it helps assess the
relative importance of each independent variable in the prediction of the
dependent variables. Secondly it identifies the magnitude and the direction
(positive or negative) of each independent variable’s relationship. Furthermore,
it helps assess the nature of the relationship between the independent
variables and the dependent variable. Finally, it provides insights into the
relationships among independent variables in their prediction of the
dependent measure.
5.3.2.2
Operationalisation of variables
The multiple regression technique was applied in the present study to
investigate the role of behavioural factors in affecting investment decisions
and to analyse the relationship between the magnitude of the investment in
renewables and the overall portfolio performance. As reported in chapter 4,
the conceptual model included three main classes of behavioural factors: a
priori-beliefs, policy preferences and technological risk attitude.
These variables were operationalised in the questionnaire using a
combination of quantitative indicators and psychometric scales (see
paragraph 5.2.1). In particular, a priori beliefs were operationalised by means
of multi-item psychometric scales. In order to increase interpretability, the
variables were then factor analyzed using orthogonal rotation. After
eliminating two items with high levels of cross loadings, the procedure yielded
a two-factor solution representing, respectively, the degree of confidence in
renewable energy technological adequacy and the degree of confidence in
market efficiency. The two variables were finally operationalised by
aggregating the items tapping into each construct. The degree of confidence
in renewable energy technological adequacy was assessed by means of the
following two items: a) energy supply from new renewable electricity sources
(e.g. wind and solar) will grow by more than 10% per year worldwide over the
next 20 years; b) solar energy is a low-density resource, requiring a lot of land:
therefore it will never achieve a significant share of the world’s energy mix
(reversed). The degree of confidence in market efficiency was assessed by
means of the following two items: c) market forces alone will never lead to a
significant exploitation of renewable (reversed); d) government intervention
does more harm than good, let governments stay out of the way.
The attitude for technological risk was assessed by means of a two-step
procedure. Respondents were first asked to allocate a hypothetical investment
budget of USD 10 Million to three different solar technologies with increasing
degrees of technological uncertainty: crystalline silicon cells, thin film cells,
70
and third-generation solar cells based on nanostructures. Technological risk
seeking attitude was then measured as the ratio of the amount allocated to
the radically innovative technologies (nanostructures) to the amount allocated
to the less innovative technologies (crystalline silicon and thin films).
The policy variables were measured using the results of the conjoint analysis.
In particular, the average importance of the attributes calculated for each
respondent was used to operationalise the policy variables retained in the
regression model. In this respect, it is worth highlighting that average
importances are calculated from part worth utilities and are expressed as the
ratio between the average utility differences between the best and worst level
of each attribute and the sum of average utilities. Since part worth utilities are
interval data which are scaled to sum to zero within each attribute, it is not
possible to directly compare values between attributes, and therefore part
worth utilities are not suited to be used as such in the regression model. As
the purpose of the regression model is to characterise the relative importance
of each policy attribute for the surveyed investors, individual-level average
importances have been used.
In order to incorporate the results of the conjoint analysis in the regression
model, a series of methodological choices had to be taken. First of all, since
the conjoint importances are expressed as percentages that add to 100, it was
not possible to include all five policy variables in the regression model, as the
latter would otherwise be undetermined. To overcome this problem, instead of
excluding one variable a priori, it was preferred to use of all policy variables by
normalizing all other utility values to the utility value of the least preferred
policy attribute, i.e. social acceptance (see chapter 7).
As mentioned, the model also included two main control variables: the
investor’s experience was measured as the number of years of experience
that each respondent had in the renewable energy sector. To control for firm’s
type, three dummy variables were created: dummy_VC included venture
capitalists and private equity firms; dummy_funds included pension funds,
hedge funds, banks and insurance companies; finally, dummy_others
included all other investors not belonging to the former two broad categories,
i.e.: project developers, utilities, infrastructure funds and engineering
companies.
The share of renewable energy in the investment portfolio was measured
through a 5-point scale, where each point corresponded to increasing
percentages of renewable energy technologies in the investment portfolio
(from 1 = less than 5% to 5 = 100%).
71
Finally, the performance of the investment was measured through a 3-point
Likert scale that assessed the extent to which the portfolio’s performance was
considered by the respondent above, equal to, or below the direct
competitor’s performance. It is worth highlighting that, although the
questionnaire included questions regarding both the past and the future
expected performance, only past performance was retained in the model as
dependent variable.
Given the confidential nature of the information collected, all the dependent
variables had to be measured through the questionnaire. This is indeed a
shortcoming of the present research. In an ideal experimental setting,
objective data should have been used to measure the investment
performance. Unfortunately, given the reluctance of the surveyed companies
to disclose information on their investment funds, more qualitative and
perceptual measures had to be adopted in the final version of the
questionnaire.
In order to exclude the risk of self reported biases and therefore confirm the
validity of results, a Harman’s single-factor test was conducted (Andersson
and Bateman, 1997; Podsakoff et al., 2003). The Harman’s single factor test
is widely used in social sciences in order to control for common method
variance (i.e. self reported bias). This refers to the amount of spurious
covariance shared among variables because of the common method used in
collecting data (Buckley et al., 1990). The test requires the researcher to load
all the variables in the study into an exploratory factor analysis and examine
the unrotated factor solution to determine the number of factors that are
necessary to account for the variance in the variables. The basic assumption
of this technique is that the presence of common method variance is indicated
by the emergence of either a single factor or a general factor accounting for
the majority of covariance among measures (Podsakoff et al., 2003). The
results of this test are reported in Table 4 and Table 5.
Table 4: Harman’s single factor test: Eigenvalues
Eigenvalue
1
2
3
4
2.95
1.72
1.26
1.00
Difference
1.22
0.46
0.26
0.20
Proportion of
variance
explained
0.33
0.19
0.14
0.11
Cumulative
variance
explained
0.33
0.52
0.66
0.77
72
As shown in Table 4, the unrotated factor solution resulting from the
exploratory factor analysis indicates a 4-factor structure with eigenvalues
greater than or equal to one and shows that none of the factors accounts for
11
the majority of the covariance among the measures . If the data were
affected by common method variance the data would have revealed a
structure largely dominated by a single factor, which should have accounted
for most of the variance in the sample.
Table 5 displays the details for the four factors identified.
Table 5: Harman’s single factor test: Unrotated factor patterns
Variables
Confidence in market
efficiency
Confidence in technology
adequacy
Technological risk
seeking attitude
Perceived importance of
the policy type
support level
support duration
the length of the
Investor’s experience
RE share in the
investment portfolio
Factor 1
Factor 2
Factor 3
Factor 4
-0.29
0.43
0.16
0.65
-0.34
0.57
0.49
0.26
-0.09
0.45
-0.74
-0.04
0.91
0.20
-0.03
0.12
0.94
0.19
0.11
0.00
0.95
0.13
0.06
0.06
0.89
0.24
0.00
-0.03
0.12
-0.54
-0.18
0.61
0.27
-0.65
-0.19
0.32
0.18
-0.56
0.48
-0.18
In the unrotated factor matrix, the columns define the factors and the rows
refer to variables. The number of factors (columns) is the number of
substantively meaningful independent (uncorrelated) patterns of relationship
among the variables (four in this case). The intersection of row and column
indicates the loading for the row variable on the column factor. The loadings
measure which variables are involved in which factor pattern and to what
degree. By comparing the factor loadings for all factors and variables, those
11
For example, factor 1 explains only 33% of the covariance among measures, factor
two is responsible for 19%, factor 3 for 14% and factor 4 explains only 11% of the
variance.
73
particular variables involved in an independent pattern can be defined, and
those variables most highly related to a pattern can also be seen. While
looking at the first factor (column), it can be seen for example that it seems to
exist a high positive correlation between the perceived importance of policy
type, support level, support duration and length of the administrative process
under factor 1 (with values of 0.91, 0.94, 0.95 and 0.89, respectively).
However, all remaining loadings under the same factor are low. Additionally,
the same variables which display high loadings under factor 1 have different
and very low values under all remaining factors. This is an additional
confirmation that the model is not affected by common method variance
problem.
5.3.2.3
Application to the present study
The conceptual model described in Figure 5 (page 51) was tested by
estimating the linear models (1) and (2) below:
Y1,i E 0 E1 x1i E 2 x2i E 3 x3i E 4 x4i E 5 x5i E 6 x6i E 7 x7i ¦ j E j control ij H i
(1)
Y2,i
J 0 J 1Y1,i J 2 x3i ¦ j J j control ij K i
(2)
Where:
Y1:
RE share in the investment portfolio
Y2:
x1 :
Confidence in market efficiency
x2 :
Confidence in technological adequacy
x3 :
Technological risk seeking attitude
x4 :
Perceived importance of policy type
x5 :
Perceived importance of support level
x6 :
Perceived importance of support duration
x7 :
Perceived importance of the length of the administrative process
Controlj: Investor’s experience, dummy VC, dummy funds, dummy ‘other
investors’
The two equations were first estimated as independent regressions by means
of Ordinary Least Squares (OLS). However, as the dependent variable of the
74
first equation (Y1) is one of the explanatory variables in the second equation,
OLS estimators could be biased and inconsistent. The model was therefore
re-estimated using three different methods: 2-stage Least Square (2SLS), 3stage Least Square (3SLS), and Seemingly Unrelated Regression (SUR) as
well. SAS Proc Model was used to estimate both the OLS and the system
equation models.
The results of a Hausman specification test (Hausman, 1978), which is used
to check for the endogeneity of a variable, suggested that the data structure
was not affected by endogeneity and that the system estimation methods
were not necessarily preferred over OLS. For consistency purposes, and to
take into account the econometric characteristics of the dependent variable,
the performance equation was also re-estimated using a multinomial logit
model. In fact, as the analysis revealed no endogeneity problems, it was
possible to estimate equations 1 and 2 separately.
The results of all the four estimation methods, as well as the additional
estimation of the performance model through multinomial logit are reported in
chapter 8.
75
6
Descriptive statistics
This section introduces the results of the questionnaire survey, by providing
some descriptive statistics for the sample analysed. More specifically, after
presenting the respondents’ profile the chapter illustrates the responses given
by the sample to the questions raised in the survey, as regards their market
and technology beliefs, the main sources informing their investment decisions,
their risk taking attitude and their self-assessment of investment performance.
This first set of results serves also a basis for the more in-depth elaborations
provided in the following chapters.
6.1
Profile of respondents
Table 6, which displays descriptive statistics, suggests that the sample is fairly
well diversified, with respect to the degree of renewable energy penetration in
the investment portfolios, the types of technologies included in the portfolios,
as well as the profile of investors.
Table 6: Descriptive statistics for the sample
Firm’s exposure to the RE investing domain
- yes
- no
Total
Investment by technology
- Solar photovoltaic
- Wind onshore
- Biomass
- Solar thermal
- CSP
- Hydropower
- Wind offshore
- Geothermal
- Biofuels
- Tidal/Wave
Research
sample
N
%
62
67%
31
33%
93
100%
N
%
36
58%
29
47%
21
34%
15
24%
14
23%
13
21%
12
19%
11
18%
4
6%
3
5%
76
Table 6: Descriptive statistics for the sample (cont.)
Share of renewables in the investment portfolio
- Less than 5%
- From 5 to 9%
- From 10 to 49%
- From 50 to 99%
- I only invest in renewables
Total
Personal experience in the RE investing domain
- No experience
- Less than 5 years
- From 5 to 10 years
- More than 10 years
Total
Company profiles
- Venture Capital, Private Equity or hybrid
- Banks, Hedge Funds, Pension Funds and
Insurance Companies
- Project developers and utilities
- Infrastructure Funds
- Private companies
- Engineering/other
- No response
Total
Age of respondents
- Under 30 years
- More than 50 years
- No response
Total
Educational background
- Economics and business administration
- Finance
- Legal
- Engineering
- Multidisciplinary
Total
N
12
6
16
11
17
62
N
10
29
17
6
62
N
34
10
%
19%
10%
26%
18%
27%
100%
%
16%
47%
27%
10%
100%
%
37%
11%
5
4
8
6
26
93
N
10
35
11
6
31
93
N
24
16
2
24
27
93
5%
4%
9%
6%
28%
100%
%
11%
37%
12%
7%
33%
100%
%
26%
17%
2%
26%
29%
100%
77
The data indicate that about two thirds of the respondents currently invest in
renewables. It is worth highlighting that the database of contacts included both
companies already investing in the renewable energy market and companies
not investing in this domain. Such distribution suggests that companies
involved in the renewable energy investing domain had a stronger interest in
taking the questionnaire compared to non-renewable energy investors.
Renewables represent at least 10% of the portfolio for over 70% of
respondents, while 27% of respondents invest only in renewables. Solar
photovoltaics and wind onshore are the two most represented technologies,
being in the investment portfolios of 57% and 47% of the respondents,
respectively. Biomass, solar thermal and concentrated solar power follow,
while tidal and wave are the least represented technologies (accounting for
only 5% of the portfolio).
With respect to the investors’ profiles, the sample is also well diversified.
Although more than half of respondents who have answered the question
declare to work for a Venture Capitalist (VC) and Private Equity (PE) Funds or
hybrid combinations of both, other investors are well represented too. Private
companies and investment funds constitute 8% of the sample. Banks,
insurance companies, pension funds and hedge funds total 10% of the
sample, project developers and utilities cover 6% of the sample, and
Infrastructure Funds account for 4%.
The average investor’s experience in the renewable energy sector is not
particularly high. Only 27% of respondents have more than 5 years of
experience with renewables, and only 10% have more than 10 years
experience. The majority (47%) declared having less than 5 years of
experience, and the remainder has not experience at all. This is not surprising,
if we consider that the renewable energy market has started experiencing a
considerable growth only quite recently and still represents a limited fraction of
the global energy market (Usher, 2008b).
It also worth adding that, although respondents seem to have rather limited
experience with the renewable energy industry, on average they have
considerable overall working experience. By looking at their age groups, it can
be noted that 84% of those who have answered this question are at least 30
years old or older, 27% are over 40 years old and 10% are older than 50
years. This suggests that investors have acquired relevant experience in other
sectors before switching to the renewable energy business. Such finding is
further confirmed by the position held within the organization. Respondents
are senior representatives of their companies, their profiles ranging from VicePresidents and CEOs to Directors, Managing Partners, Associates and Senior
78
Analysts. The high-level profile of investors ensures a certain level of
accuracy in the responses, thus supporting the reliability of data.
Finally, in terms of education almost one third of respondents have a
multidisciplinary background; 25% have studied Engineering and 25%
Economics and Business Administration.
6.2
Awareness and beliefs
As mentioned in chapter 5, the purpose of the second section of the
questionnaire is to assess the investors’ knowledge and awareness of the
technological and market potential of renewable energy sources, their a-priori
beliefs about the role of market and policy in supporting the growth of
renewables, as well as their attitude toward technological innovation.
The first question aims at investigating the degree of awareness regarding the
cumulative amount of investments needed in the energy sector in order to
invert the current trend of increasing greenhouse gas emissions and avoid
catastrophic consequences for the climate.
The background data used to formulate the question are taken from the
already cited IEA publication “Energy Technology Perspectives” (ETP), 2008
edition. The ETP represents a reference book for energy practitioners, and is
widely known also to most renewable energy investors and commercial
organisations. Data refer to two different CO2 mitigation scenarios: the “ACT
scenario”, resulting in a stabilization of CO2 emissions to current levels by
2050, and the “BLUE MAP scenario”, which would lead to a halving of CO 2
emissions by the same target year.
The question addressed to the survey respondents was: “If CO 2 emissions are
to be significantly reduced by 2050 compared to current levels, huge
investments in new and innovative technologies are required. How would you
assess the following estimates of investment needs provided by various
researchers”.
The responses given by the investors in the first case are displayed in Figure
9.
79
Figure 9: Respondents’ opinions on the appropriateness of the
investment effort proposed by the ACT scenario of the ETP 2008
publication
For almost half of respondents, an additional investment effort of 17 trillion
over the period 2005-2050 is not enough to avoid serious consequences for
the climate. This suggests that on average investors seem to be quite aware
of the magnitude of the problem. Another 18% of the sample believes that the
proposed amount is appropriate, while only 3% of investors think this effort is
too high.
Figure 10 shows the responses provided by the investors regarding the more
ambitious emission reduction scenario. An increase in the global investment
needs of USD 45 trillion between 2005 and 2050 compared to a business as
usual scenario is considered appropriate by over 37% of respondents. Almost
one fourth of the sample considers this estimate too high, and more than 12%
of respondents believe it is too low.
80
Figure 10: Respondents’ opinions on the appropriateness of the
investment effort proposed by the BLUE MAP scenario of the ETP 2008
publication
In the following question investors are asked to identify the likely share that
renewables could reach in Europe by 2050.
As displayed in Figure 11, a penetration rate of 20% of renewables by 2050 is
feasible for almost 80% of respondents. In addition, over 40% of the sample
believes that renewable energy sources could reach a rate of 50% by 2050. A
higher share of renewables in the future European energy mix is not likely to
be reached, according to the investigated sample. In particular, only 9% of
respondents believe that reaching up to 80% of renewable energy in the
European mix by 2050 is feasible, while 62% think that this target is
unfeasible. Furthermore, 74% of respondents think that reaching a fully
renewable-based energy system is completely unfeasible.
81
Figure 11: Respondents’ opinions on the feasible share that renewables
could reach in Europe by 2050
The third question consists in a table displaying seven common beliefs about
renewable energy sources. Respondents are asked to indicate to what extent
they would agree with each of the statements proposed.
These a priori beliefs can be grouped in two main categories: technology
beliefs, which reflect the investors’ trust in the potential of the technologies
considered for the investment, and market and policy beliefs, which are
related to the investors’ degree of confidence in the efficiency of market and
policy mechanisms. As already mentioned, following a cognitive psychology
approach some statements are positively phrased while some others are
negatively phrased.
As shown in Figure 12, two thirds of respondents believe that the energy
supply from new renewable energy technologies like solar and wind will grow
by more than 10% per year worldwide over the next 20 years. This is in line
with the projections made available by the most authoritative sources in the
sector (IEA, 2008a) and corroborates the impression that, on average,
respondents display a high degree of awareness and a good knowledge of the
technology potentials.
82
Figure 12: Respondents’ opinions on the growth potential of new
renewable energy technologies
Figure 13 reports one common belief about solar energy, i.e. the fact that it
will never achieve a significant share of the world’s energy supply given the
huge amount of land that it requires.
Figure 13: Respondents’ opinions on the potential for solar energy to
reach a significant share of the world’s energy mix
This common misperception does not take into account the significant
progresses achieved by solar technology, which makes it suitable for buildingintegrated applications, thus not requiring any land occupation. Respondents
seem to be aware of these technological improvements; in fact over 60% of
83
them do not agree with this statement. However, it is worth highlighting that
14% of investors either agree or strongly agree.
Another common opinion about renewable energies is that their intermittent
availability raises a lot of additional problems to grid management since it
alters grid stability. It is believed that huge investments are needed in order to
ensure storage systems and back-up capacities to compensate for the
variability of renewable energy sources. In reality, the amount of renewables
that can be safely integrated in grids depends on the flexibility of the whole
system. The latter depends on four aspects: flexible supply (for instance
based on gas and hydro), flexible demand (e.g.: smart grids), storage and
interconnections. In particular, a highly interconnected system needs much
less storage and back-up. Since huge investments in transmission and
distribution are needed in a scenario of highly growing electricity demand, this
is a unique opportunity to modernise and make electricity systems more
flexible, which will allow for an easier integration of renewables.
Although significant improvements have been reached in terms of flexibility of
power systems and of renewable energy output forecasts, many stakeholders
are still skeptical on the actual potential of renewables for large scale
deployment because of the mentioned shortcomings. As far as the surveyed
investors are concerned, Figure 14 indicates that almost half of respondents
do not agree with this statement, while 28% either agree or strongly agree.
Figure 14: Respondents’ opinions on the large-scale growth potential of
renewable energy technologies given their intermittent availability
As already described in chapter 2, experts indicate that with current trends
solar PV will reach the so called grid-parity within the next 10 to 20 years.
84
Therefore, in order to measure the degree of optimism of investors regarding
the potential of solar PV, a specific question has been elaborated on this topic.
As Figure 15 reveals, this is a quite controversial issue among the investment
community; in fact the number of investors who do not believe this will happen
almost equals the number of those who believe that solar PV will reach the
same cost of conventional retail electricity within the next decade (41% versus
46%).
Figure 15: Respondents’ opinions on the potential of PV to reach grid
parity
As far as policy and market beliefs are concerned, three specific questions
were addressed in the survey, which have the purpose to understand whether
respondents see policies and regulations as beneficial or harmful to the
effective deployment of a market for renewables.
According to Figure 16, 44% of respondents think that government regulation
is important to enable a favourable framework for the growth of renewable
energies, and that market forces alone will never be able to boost the sector.
However, 35% of the sample has the opposite view, i.e. that governments
should not interfere with market dynamics, and that market forces will ensure
the significant exploitation of renewable energy sources.
The statement presented in Figure 17 presents a completely opposite view
compared to the previous one, since it incorporates an anti-government
perspective expressed by one venture capitalist during a previous survey
(Teppo, 2006).
85
The results indicate that for over 70% of respondents, government
intervention in the renewable energy sector is not harmful; only 8% of the
sample concurs with the proposed statement.
Figure 16: Respondents’ opinions on the possibility for market forces to
lead to a significant exploitation of renewables
Figure 17: Respondents’ opinions on the appropriateness
government to intervene in the renewable energy sector
for
The last statement intends to provide support to one of the previous two, by
looking at investors’ attitudes vis-à-vis renewable energy subsidies. This topic
is very controversial and has been extensively disputed by both scholars and
practitioners. Opponents point out that renewable energy subsidies distort the
86
market therefore leading to unfair competition, while supporters highlight that
subsidies to renewable energies are just a way to create a more level playing
field. In fact, fossil energy sources are also heavily subsidized by
governments, with estimated annual expenditures of approximately 550 billion
USD (Birol, 2010).
According to fifty percent of the investigated sample, renewable energy
subsidies do not represent a potential source of risk for their investments.
Another 23% of respondents on the contrary share the view that subsidies
represent a source of risk, as shown in Figure 18.
Figure 18: Respondents’ opinions on the appropriateness of renewable
energy subsidies
6.2.1
Correlations among a priori beliefs
Once described qualitatively the main investors’ technological and market a
priori beliefs, an additional analysis has been carried out, which consisted in
looking for some correlation amongst these beliefs. The contingency table
reported below (Table 7) reveals a strong correlation (significance at the 99%
level) between some of them. In particular:
- Those who think that electricity generated by solar PV will reach grid parity
within the next decade do not support the view that solar energy will never
achieve a significant share of the world’s energy mix, thus displaying a strong
confidence in the potential of solar technology
- Those who think that electricity generated by solar PV will reach grid parity
within the next decade believe also that energy from new renewable energy
sources will grow by over 10% per year worldwide over the next 20 years,
87
thus confirming their positive attitude vis-à-vis the potential of more innovative
technologies
- Those who think that solar energy will never achieve a significant share of
the world’s energy mix believe also that large scale deployment of renewables
is severely hampered by their intermittent variability, thus implying that these
investors still perceive high technological barriers; for them achieving a
satisfactory level of technological reliability is an important condition to commit
more capital in the renewable energy business sector.
88
Table 7: Significant correlations between a priori beliefs
89
6.2.2
Technology beliefs and share of investments in renewables
The purpose of this analysis was to understand whether a higher degree of
confidence in the technological potential of renewable energy sources is
reflected into a higher share of renewables in the investment portfolio. In the
present paragraph the main findings are reported, according to the specific
types of technological a priori beliefs analysed.
From the analysis of contingency tables reported below (Table 8- Table 10) it
can be seen that:
- The stronger is the belief that new renewable energy technologies will grow
by 10% per year worldwide over the next 20 years, the higher is the share of
renewables in the investment portfolio
- The stronger is the belief that solar will never achieve a significant share of
the world’s energy mix, the lower is the share of renewables in the investment
portfolio
- The stronger is the belief that large scale deployment of renewables is
severely hampered by their intermittent availability, the lower is the share of
renewables in the investment portfolio
90
Table 8: Analysis of the contingency table comparing the opinions on the growth potential of new renewable
energy technologies with the actual share of RES in the investment portfolio
Energy supply from new renewable electricity sources (e.g. wind and solar)
will grow by more than 10% per year worldwide over the next 20 years
N=77
What
percentage of
your portfolio
is in
renewable
energy
technologies?
Total
I strongly
disagree
I disagree
I am
indifferent
I agree
I strongly
agree
Total
Less than 5%
10.0%
10.0%
20.0%
40.0%
20.0%
100.0%
From 5 to 9%
0.0%
0.0%
50.0%
50.0%
0.0%
100.0%
From 10 to
49%
0.0%
8.3%
8.3%
66.7%
16.7%
100.0%
From 50 to
99%
0.0%
0.0%
12.5%
25.0%
62.5%
100.0%
I only invest
in RES
0.0%
7.1%
0.0%
71.4%
21.4%
100.0%
2.1%
6.3%
12.5%
54.2%
25.0%
100.0%
91
Table 9: Analysis of the contingency table comparing the opinions on the growth potential of solar energy
technologies with the actual share of RES in the investment portfolio
Solar energy is a low-density resource, requiring a lot of land. Therefore it
will never achieve a significant share of the world’s energy mix
N=77
What
percentage of
your portfolio
is in
renewable
energy
technologies?
Total
I strongly
disagree
I disagree
I am
indifferent
I agree
I strongly
agree
Less than 5%
10.0%
30.0%
From 5 to 9%
0.0%
25.0%
From 10 to
49%
33.3%
From 50 to
99%
I only invest
in RES
Total
50.0%
0.0%
10.0%
100.0%
50.0%
25.0%
0.0%
100.0%
50.0%
0.0%
16.7%
0.0%
100.0%
50.0%
25.0%
25.0%
0.0%
0.0%
100.0%
28.6%
57.1%
7.1%
7.1%
0.0%
100.0%
27.1%
41.7%
20.8%
8.3%
2.1%
100.0%
92
Table 10: Analysis of the contingency table comparing the opinions on the large-scale growth potential of
renewable energy technologies with the actual share of RES in the investment portfolio
Large-scale deployment of renewables is severely hampered by their
intermittent availability
N=77
What
percentage of
your portfolio
is in
renewable
energy
technologies?
Total
I strongly
disagree
I disagree
I am
indifferent
I agree
I strongly
agree
Total
Less than 5%
0.0%
20.0%
30.0%
40.0%
10.0%
100.0%
From 5 to 9%
0.0%
0.0%
50.0%
50.0%
0.0%
100.0%
From 10 to
49%
16.7%
33.3%
33.3%
8.3%
8.3%
100.0%
From 50 to
99%
25.0%
50.0%
12.5%
12.5%
0.0%
100.0%
I only invest
in RES
14.3%
57.1%
14.3%
14.3%
0.0%
100.0%
12.5%
37.5%
25.0%
20.8%
4.2%
100.0%
93
6.3
Main sources for informing investment decisions
The questionnaire also intended to provide some useful indications regarding
the most relevant sources to inform investment decisions in the renewable
energy market. To this end respondents were asked to rank the following five
sources based on the extent to which they influence their investment
decisions: i) investments by well-known/high-profile investors in the sector; ii)
technical reports; iii) personal intuition; iv) in-house due diligence; v)
consultants' opinion.
These options were presented by means of a pull-down list and measured
along a one-to-five scale, where 1 corresponds to the most important source
and 5 to the least important source.
Figure 19 shows that in-house due diligence scores highest amongst the
proposed options, being mentioned as the most important source of
information by 35% of respondents. Financiers seem to rely also on the
investment decisions taken by sector leaders (25% of responses), and on
technical reports (22%). Personal intuition is mentioned by 11% of
respondents, while consultants’ opinion is the least relevant information
source according to respondents.
Figure 19: Most important sources for informing investment decisions
(items having received a 1 ranking in the survey; N=65)
In order to gain even deeper insights, some demographic segmentation has
been conducted.
94
Table 11 reveals that young professionals seem to display a certain
overconfident attitude, which is progressively softening with the increase of
experience. For example, while personal intuition is mentioned as the most
important driver for an investment decision by 50% of young financiers, very
experienced professionals (more than 50 years old) declare to rely mainly on
in-house due diligence.
Intermediate groups (31-40 and 41-50 years old investors) have a more
diversified attitude. Among the various categories, they tend to follow more
what well known and highly reputed investors are doing, while taking their
investment decisions.
Table 11: Most important sources for informing investment decisions
split by age groups
Information sources
(N=65)
Age group
31-40
41-50
38%
43%
22%
60%
0%
21%
44%
20%
Technical reports
13%
14%
33%
20%
My personal intuition
50%
11%
0%
0%
Consultants’ opinion
0%
11%
0%
0%
100%
100%
100%
100%
In-house due diligence
Investments by well known/high
profile investors
Total
< 30
> 50
Results split by type of company are more nuanced. As shown in Table 12,
there is no specific behaviour that characterises strongly a certain company
type compared to others. Similar patterns can be identified in pension funds
and hedge funds. However, the sample of these two categories is too small to
enable any strong conclusion. The other interesting aspect is that VC and PE
funds assign the highest rankings to personal intuition across the whole
sample.
95
Table 12: Most important sources for informing investment decisions
split by company type
Information
sources
1
(N=65)
In-house
due
100%
diligence
Investments
by well
0%
known/high
profile
investors
Technical
0%
reports
My personal
0%
intuition
Consultants’
0%
opinion
Total
100%
1= Pension Fund
2= Venture Capital
3= Bank
4= Hedge Fund
5= Insurance
6= Private Equity
7= Other
2
3
Company type
4
5
6
7
0%
20%
100%
0%
21%
30%
8%
20%
0%
100%
14%
30%
17%
20%
0%
0%
21%
5%
58%
20%
0%
0%
36%
35%
17%
20%
0%
0%
7%
0%
100%
100%
100%
100%
100%
100%
Figure 20 regroups the results for all scores and for the whole sample.
Figure 20: Main sources for informing investment decisions (from
1=most important to 5=least important)
96
By summing up the scores received by each information source as first and
second preferred option respectively, technical reports outstrip in-house due
diligence as an important source for informing the investment decision making
process. It is also worth highlighting that investments performed by wellknown market actors are mentioned as an important source of information by
40% of the sample, thus confirming the existence of herd behaviours in this
market. Personal intuition is an important source of information for over one
third of investors, while consultants’ opinions remain by far the least
considered source by the investigated sample. Finally, when looking at
weighted average scores the following ranking order is obtained: 1) in-house
due diligence and technical reports, 2) investments by well known/high profile
investors in the sector, 3) my personal intuition, and 4) consultants’ opinion.
6.4
Risk taking attitude of investors
In order to evaluate the investors’ attitude for technological risk, a question
was formulated, which asked respondents to allocate a hypothetical
investment budget of USD 10 Million to three different solar technologies with
increasing degrees of technological uncertainty (and therefore with increasing
degrees of risk): crystalline silicon cells, thin film cells, and third-generation
solar cells based on nanostructures. The results are presented in Figure 21,
which shows the average allocation for the sample: 3.7 million Euros would be
hypothetically invested in crystalline silicon solar cells, 3.7 million Euros in
thin-films solar cells, and 2.6 million Euros in third generation solar cells.
Figure 21: Distribution of the portfolio amongst PV solar technologies
97
This average distribution appears quite rational, since it takes accurately into
account the current market share of the various solar technologies and their
expected evolution. Indeed, as reported by EPIA (2010) while the PV market
is still dominated by crystalline silicon technologies, more innovative
technologies like for example thin films are expected to gain a higher share in
the next few years, thanks to the more accelerated compound annual growth
rate they are currently displaying.
However, by looking more in detail at single responses, quite significant
differences can be identified among respondents. In particular, 20% of the
sample would not invest any Euro on crystalline silicon solar cells, 12% would
not invest on thin-films at all and 18.5% would not put any money on third
generation solar cells. At the other extreme, 6% of respondents show a high
risk aversion, since they would entirely invest on the most mature technology,
which is represented by crystalline silicon solar cells. Another 6% would
allocate the money in thin-films, a technology which is progressively gaining
market share – nearly 23% in 2009 - but that is still significantly less exploited
than crystalline silicon (REN21, 2010). Finally, 3% would invest only on thirdgeneration solar cells, thus displaying a particular risk-taking attitude.
In order to better understand the degree of awareness of investors in this
specific renewable energy market segment and to get a better feeling on their
risk profile, a comparison was made between the average survey responses
and the likely evolution of the PV market distribution in 2030, as reported in
the IEA ETP publication (IEA, 2008a). The latter reports the projections to
2050, based on a comprehensive analysis of the historical evolution of costs
and learning curves, market growth, technology developments and policy
targets, as well as the review of the most authoritative technology and policy
roadmaps, and the discussion of the main potential and barriers of PV
technologies. As such, it can be used as an objective reference for the likely
evolution of the PV market in the next decades.
Although the time horizons do not necessarily coincide (in particular,
investment decisions may have a shorter horizon than 2030), it is remarkable
to see that on average the distribution resulting from the survey perfectly
matches the expected market share of PV technologies in twenty years from
now (see Figure 22). This confirms that survey respondents have a quite
robust knowledge of the current status and the expected evolution of
renewable energy technologies, thus basing their decisions on factual
considerations
98
Figure 22: Forward-looking attitude of investors
6.5
Assessment of performance
In the fourth section of the questionnaire, respondents were solicited to
provide a self-assessment of the perceived past performance and of the
expected future performance of their investments compared to their direct
competitors. Results are displayed in Figure 23 and Figure 24, respectively.
As far as past performance is concerned, based on the three best performing
investments done over the past five years, two thirds of respondents believe
that their performance is in line with that of their direct competitors. More than
one third of respondents think that their performance is well above the
performance of direct competitors, whereas about 6% of respondents think
that they underperform their direct competitors.
Looking at the expected future investment performance, over half of the
sample has an optimistic view, thinking that the portfolio’s return will be higher
or significantly higher compared to current levels. Over 44% of respondents
believe that their performance will remain in line with current levels, and only
4% of respondents think that future performance will be lower than current
levels.
99
Figure 23: Perceived past investment performance compared to direct
competitors
Figure 24: Expected future investment performance compared to direct
competitors
Also in this case, some demographic segmentation has been conducted in
order to see whether there is any influence of age and type of company on the
perceived future performance.
Results are shown in Table 13 and Table 14. Regardless of the age, all
investors display a positive attitude, which is even more pronounced in more
experienced professionals. There is also no significant difference amongst
investors’ profiles. All types of companies in fact think that in one year from
100
the investment decision, the performance of their investment will be at least in
line with current levels, and in many cases higher.
Table 13: Perceived future investment performance, by age groups
Perceived future investment
performance
(N=62)
Age group
< 30
31-40
41-50
> 50
Lower than current levels
12%
8%
0%
0%
In line with current levels
38%
46%
33%
60%
Higher/significantly higher than
current levels
50%
46%
67%
40%
100%
100%
100%
100%
Total
Table 14: Perceived future investment performance, by company type
Perceived future
investment
performance
(N=66)
Company type
1
Lower than current
levels
2
3
4
5
6
7
100%
0%
0%
0%
0%
7%
0%
In line with current
levels
0%
55%
40%
0%
0%
50%
39%
Higher/significantly
higher than current
levels
0%
45%
60%
100%
100%
43%
61%
100%
100%
100%
100%
100%
100%
100%
Total
1= Pension Fund
2= Venture Capital
3= Bank
4= Hedge Fund
5= Insurance
6= Private Equity
7= Other
101
7
Analysis of policy preferences
As already described, an important step in the research process was the
assessment of investors’ attitudes toward renewable energy policies. Since
renewable energy markets are regulated under several policy schemes,
understanding which policy characteristics influence the most the likelihood to
invest in a renewable energy project by reducing the perceived risk associated
to the investment can provide useful insights to decision makers. To achieve
this goal, adaptive conjoint analysis was used as a tool to investigate
investors’ preferences over renewable energy policy characteristics.
Below the main findings are presented and discussed. The results have been
calculated using the Hierarchical Bayes approach, which is considered to
provide more accurate estimates compared to conventional method for
adaptive conjoint analysis like Ordinary Least Squares (Sawtooth, 2002 and
2006). However, in order to assess the influence of the evaluation method on
results, a sensitivity analysis has been conducted using Ordinary Least
Squares technique. The results are reported in paragraph 7.7.
7.1
General findings for the sample
The results of the adaptive conjoint analysis indicate in a clearer way than the
descriptive analysis that the sample is composed of rather risk-averse
investors who seek to maximize their return over a relatively short time
horizon. Indeed, the most relevant policy attribute is the level of the premium
incentive, which on average receives over 25% of preferences. The type of
policy scheme is ranked second in terms of importance, being assigned over
21% of preferences. The third important element of policies according to the
investigated sample is the duration of the financial support given to the
renewable energy project. This seems to be almost equally important as the
type of policy scheme, since it receives about 21% of preferences from the
respondents. The length of the administrative process is ranked fourth,
receiving 18% of preferences.
The least important policy attribute for the sample is the overall degree of
acceptance toward the renewable energy technology manifested by the main
stakeholders in the hypothetical policy framework described in the survey. In
fact, social acceptance receives slightly more than 14% of preferences.
The above-described results are depicted in Figure 25, and in Table 15, which
reports also the standard deviations.
102
Figure 25: Average importance of policy attributes for the investigated
sample
Table 15: Average importance of policy attributes for the investigated
sample, and standard deviations
Policy attribute
Average importance
Standard deviation
Level of premium
incentive
25.22%
4.32
Type of RE support
scheme
21.48%
5.07
Duration of the support
20.91%
3.98
Length of the
18.08%
4.38
Social acceptance
14.31%
5.15
100.00%
(N=60)
The information provided above represents a first interesting result of the
analysis. However, since data are presented in an aggregated way they do
not provide enough detail on the different utilities associated to varying levels
of the attributes described.
These additional details are provided in Figure 26 and Table 16, which show
the part worth utilities. Part worth utilities describe the contribution of the
various attribute levels to the overall utility, thus allowing to understand how
the change in a variable level affects the investor’s preferences for a certain
103
policy framework. Part worth utility results are normalized using the Zerocentered differentials (diffs in the following) method. The diffs method rescales
utilities so that the sum of the utility differences between the worst and the
best level of each attribute is equal to the number of attributes times 100
(Orme, 2009b). This normalizes the data so that each respondent has equal
impact when computing the population average.
N=60
Figure 26: Importance of the various attribute levels
104
Table 16: Average utilities of policy attribute levels for the investigated
sample, and standard deviations (N=60)
Zero-centered
Average utilities
Level of the premium incentive
100 €/MWh
60.21
75 €/MWh
23.60
50 €/MWh
-17.91
25 €/MWh
-65.91
RE support scheme:
Feed-in tariffs
57.35
TGC/RPS
14.57
Tender schemes
-21.85
Tax incentives/
-50.06
Investment grants
Duration of the support:
More than 20 years
51.21
10-20 years
2.13
Less than 10 years
-53.35
Length of the administrative process
<6 months
44.45
6-12 months
1.49
>12 months
-45.93
Social acceptance:
High
35.77
Low
-35.77
Standard deviation
10.11
7.33
4.17
15.67
13.98
4.72
6.63
8.22
10.20
5.10
10.32
10.73
4.36
11.58
12.88
-12.88
This additional information reveals that significant differences exist also within
each attribute. Starting from the most important policy attribute, i.e. the level of
support granted to the renewable energy project, one can see that a premium
incentive of 100 €/MWh is more desirable than a premium incentive of 75
€/MWh (60.21 against 23.60). This confirms the finding that the sample is
composed of agents who look for the highest return of their investment. It is
worth highlighting that the negative values associated to some attribute levels
do not imply that these options are not desirable per se. In fact, by applying
the Zero-centered diffs method, the utilities have been re-scaled to an
arbitrary additive constant to sum 0 within each attribute. As a consequence, a
negative value of part worth utilities does not mean that the particular attribute
level is unattractive in absolute terms, but only that it is less preferred than an
alternative option showing a higher part worth utility value. In other terms,
105
while positive values indicate an increase in the utility, negative values
indicate a decrease in utility for the observed respondent.
Regarding the type of renewable energy support scheme, feed-in tariffs are by
far the most preferred instrument, with an average utility of 57.35. More
market-based support mechanisms such as tradable green certificates or the
renewable portfolio standard, receive an individual utility of 14.57. This finding
is in line with a previous research conducted by Bürer (2008), where feed-in
tariffs were found to be the most favoured renewable energy support scheme
by a group of European and US venture capital and private equity investors. A
following research conducted by Sovacool via interviews with 181 energy
experts worldwide (Sovacool, 2009) confirms that feed-in tariffs are a very
popular scheme, ranking third among the most favoured policy options
identified by participants while tradable green certificates rank eleventh.
It is also worth adding that currently feed-in tariffs are the most diffused
support instrument in the EU, with 21 countries out of 27 having in place feedin tariff as either a partial or exclusive way to support renewables (Pfluger,
2009). Although all renewable energy support scheme options were described
in detail in the survey in order to induce a perfectly informed choice, one
cannot exclude that the very high ranking assigned by investors to feed-in
tariffs might have been - at least partly - influenced by the popularity of this
instrument compared to less familiar alternatives.
In their turn, tradable green certificates have been considered for long time as
the most effective and cost-efficient means to stimulate renewable electricity
production (EC, 1999; Verhaegen et al., 2009). More recently however, their
popularity has started going down. This is reflected in the more reduced
number of countries currently adopting this scheme in the EU (among them:
the United Kingdom, Sweden, Belgium and Italy).
Tender schemes for renewables, as applied for example in Denmark, France,
Ireland and the United Kingdom, are ranked third, with a utility value of -21.85.
Finally, fiscal measures like tax incentives of investment grants receive by far
the lowest utility value amongst the proposed options (-50.06). Again, this
result might be partly explained by a certain lack of confidence with this
instrument by the surveyed investors. In fact, this type of support scheme is
currently in place only in a very limited set of EU Member States, like Finland
and Malta, which were outside the scope of the survey.
In terms of the duration of the support granted to the renewable energy project,
investors show a high interest for a long-term support: an incentive paid for
more than 20 years is strongly preferred over an incentive paid for a 10 to 20
106
years timeframe (51.21 compared to 2.13, respectively). A time horizon
shorter than 10 years receives a very negative score (-53.35), meaning that in
order to embark in a renewable energy investment, agents ask policy makers
to guarantee a financial support to the project over a minimum timeframe of
10 years.
As far as the duration of the administrative process is concerned, investors
look for a smooth framework allowing to get the necessary authorizations and
permits in a relatively short timeframe. A quick administrative process (less
than 6 months) receives an average individual utility of 44.45, a medium
process (6-12 months) records an individual utility of 1.49 while a slow
administrative process (requiring more than 12 months) is assigned an
individual utility of -45.93. This can be interpreted as a percentage change in
choice probability: if all other elements of the policy framework remain
constant, the likelihood that an investor prefers a given policy framework
increases by approximately 50% if the administrative process is shortened
from more than 12 months to 6-12 months.
The relatively lower importance assigned by the surveyed investors to the
duration of the administrative process seems to contradict some preliminary
findings from Lüthi and Wüstenhagen (2009), who found this policy attribute to
be the most relevant for the investigated sample. However, this difference can
be explained by at least three factors: i) the different sample analyzed (a
group of project developers against a more diversified profile of investors); ii)
the different geographical context (a selected group of mainly South EU
countries against a broader group of EU countries); and, iii) the different
operationalisation of the variable (5 levels of the attribute instead of the three
levels used in the present survey).
As for the first factor, it is to presume that the length of the administrative and
authorization process is a serious issue of concern for project developers,
whereas this is certainly not an issue for other category of investors like for
example insurance companies, banks or venture capital funds. In this respect,
a segmentation analysis has been conducted (see paragraph 7.3), which has
revealed that the company type has some influence on investors’ preferences
over different policy attributes. However, results also suggest that the
geographical location where the investor operates might be a better
parameter to explain the different values assigned to attribute levels by similar
companies.
Regarding the second factor, several studies have already pointed out that
administrative hurdles represent a barrier to renewable energy investments in
some EU countries like for example Greece, the Netherlands, Spain and the
107
United Kingdom (Papadopoulos and Karteris, 2009; Reiche and Bechberger,
2004), while they are less of an issue in other countries like Germany or
Finland. One of the major conclusions of the analysis of barriers to the
development of renewable electricity in the EU Member States carried out
within the OPTRES study (OPTRES, 2006) is that in general administrative
and regulatory hurdles are perceived to be a quite severe obstacle in the
development of renewable energy projects, according to project developers
and other stakeholders. However, the magnitude of the obstacle is perceived
differently by stakeholders, depending on the type of renewable energy
technology and the national situation. The latter issue is particularly true for
grid barriers.
A
recent
study
developed
by
the
Windbarriers
consortium
(http://www.windbarriers.eu; cited in Ragwitz, 2010) has revealed that actually
administrative and grid-related barriers show a large variety in the EU Member
States. The shortest lead time for the authorization process is found in Finland,
followed by Belgium. Quite surprisingly, Germany shows a high variability in
the lead time and an average length of the process higher than other EU
countries, like for example Italy. However, this can be due to the fact that
results are averaged across different types of technologies, which are
subjected to different authorization processes. For example, in Germany the
procedures related to offshore wind projects are still complicated compared to
other technologies (OPTRES, 2006), and this might have affected the overall
results.
Finally, from the results of Lüthi and Wüstenhagen (2009), it seems that
project developers are particularly attracted by the possibility to get the
authorization in 1- 2 months timeframe; anytime longer than this time span
becomes less preferable compared to other alternatives. Since in the current
survey a different measurement scale was adopted, it is not possible to
capture this difference. However, it is worth highlighting that according to
previous surveys (OPTRES, 2006) on average acquiring permits for grid
connection can take quite a long time, which ranges from 6 to 16 months in
the case of France to up to 10 years in the case of Cataluña.
With respect to the last policy attribute analyzed, a higher social acceptance is
obviously preferred to a more resilient framework for renewable energy
investments. Nevertheless, the social acceptance component is less relevant
compared to the other policy attributes previously analyzed. This finding can
be interpreted in several ways. One explanation could be that investors
believe that addressing social acceptance issues is a task to be carried out by
policy makers, not by private operators. This is tantamount to saying that
108
investors look for adequately designed top-down policies that take into
account all relevant barriers that might hamper the diffusion of renewable
energy technologies, including social acceptance.
Another possible explanation is that investors believe that onshore wind
technology is a relatively not disputed technology. Therefore they tend to
assign a relatively lower importance value to social acceptance simply
because they do not perceive a strong opposition to this renewable energy
technology in the EU community. Previous research seems to support this
explanation. In fact, the results of the OPTRES survey (OPTRES, 2006)
indicate that onshore wind is the second least controversial renewable energy
technology after solar photovoltaics, in the perception of investors.
However, again important national differences can be observed. For example,
social acceptance scores lowest in countries like Ireland, Estonia, Italy and
France, whereas it receives a high ranking (meaning high barrier) in Belgium,
Slovenia, the Netherlands and Austria, mostly for aesthetic or noise reasons.
Survey respondents indicated that the involvement of local population in the
planning phase can significantly decrease opposition, and therefore
recommended clear information and early participation in the decision making
process. Allocating revenues to local authorities or making them financial
partners in the development of the project are two suggested solutions in
order to overcome social acceptance barriers.
Reiche and Bechberger (2002) also suggest that consensus-based decision
making is a very important factor in order to increase the share of renewables.
In this respect, policy can influence public awareness, for example by granting
tax exemptions for renewable energy.
Finally, it is worth remembering that a large majority of the surveyed sample is
represented by venture capital and private equity firms, which operate in the
early stage of the project value chain. These actors typically invest in
technology assets, but do not physically build renewable energy projects.
Therefore they are less concerned by social acceptance issues than other
investors’ typologies like for example project developers and utilities. In order
to account for the variation on these perceptions according to the different
profiles of investors, a sensitivity analysis has been conducted (see paragraph
7.3 below).
109
7.2
Analysis of single policy attributes
After presenting the data in an aggregated way for the total of the investigated
sample, in the following the various policy attributes are further analyzed in
order to show the distribution of individual utilities.
7.2.1
The highest level of the premium incentive presented to the investors during
the survey was Euro 100/MWh. The raw utility assigned to this attribute level
12
ranges from a minimum of 1.27 to a maximum of 2.78, with a mode of 1.27 ,
a mean value of 2.14 and a median value of 2.2. It is worth remembering that
in a symmetric distribution these common measures of central tendency are
identical, while differences occur in case of skewed distributions.
As one can see from Figure 27, the frequencies are quite symmetrically
distributed around the mean value. This is confirmed by the low value of the
standard deviation (0.39), indicating that the data points tend to be very close
to the mean. About 68% of the values fall within one standard deviation of the
mean (2.14+-0.39). The values of the 25° and 75° percentiles are,
respectively, 1.94 and 2.41. The asymmetry coefficient is -0.526 while kurtosis
is -0.462, meaning that the distribution is slightly skewed to the left.
Figure 27: Distribution of raw utilities for 100 €/MWh
12
Actually, this is a multimodal distribution. Only the lowest mode is reported here.
110
Figure 28 shows the distribution of individual utilities for the second level of
premium incentive proposed: 75 €/MWh.
The minimum level observed is 0.37 and the maximum is 1.56. The mean and
median values are quite close (0.84 and 0.81, respectively) while also in this
case more modes are observed (the minimum value is 0.37). The standard
deviation has a value of 0.27. Approximately 69% of the frequencies are
comprised between one standard deviation of the mean. The values of the 25°
and 75° percentiles are 0.63 and 1.04. The asymmetry value is 0.35 and the
kurtosis is -1.15.
The distribution of individual utilities for the level of incentive 50 €/MWh is
shown in Figure 29. The minimum and maximum levels observed are,
respectively, -0.97 and -0.28. The mean value for this distribution is -0.63,
while the median value is -0.66.
Multiple modes are observed, with a minimum value of -0.93. The standard
deviation has a value of 0.15, meaning that the frequency distribution tends to
be very close to the mean. In fact, 72% of the frequencies are comprised
between one standard deviation of the mean. The values of the 25° and 75°
percentiles are -0.74 and -0.52. The asymmetry value is 0.24, and the kurtosis
is -0.25. These values confirm that the observations are quite normally
111
distributed, although slightly skewed to the right and that the observations are
not dispersed along the tails.
The last level of the premium incentive analysed (25 €/MWh) has a mean
value of -2.35 and a median value of -2.49. The minimum and maximum
levels observed are, respectively, -3.37 and -1.17. The standard deviation has
a value of 0.57. This implies that the frequency values are more dispersed
compared to the previous distributions, as one can see from Figure 30.
The analysis of frequencies reveals that 62% of data are comprised between
one standard deviation of the mean. The values of the 25° and 75° percentiles
are -2.84 and -1.91. The asymmetry value is 0.28, and the kurtosis is -0.93,
indicating a platykurtic distribution characterized by a wider peak around the
mean and thinner tails.
112
7.2.2
Type of renewable energy support scheme
As reported above, in the survey four different support schemes were
proposed to respondents: feed-in tariffs, tradable green certificates/renewable
energy portfolio standards, tender schemes and tax incentives/investment
grants.
Figure 31 illustrates the distribution of raw utilities calculated with the
Hierarchical Bayes method for the type of scheme “feed-in tariffs”. This
attribute level was assigned the highest share of preferences by respondents.
The raw utilities range from a minimum value of 1.14 to a maximum of 2.98.
The values of the mean and the median are quite close (2.04 and 2.06,
respectively). The minimum value reached by the modes is 1.14. The values
of the 25° and 75° percentiles are 1.56 and 2.38. The graph shows a
platykurtic distribution, characterized by a wider peak around the mean and
thinner tails. This is confirmed by the relatively higher value of the kurtosis (1.05) compared to the other observations previously shown. The asymmetry
is 0.03, meaning that the distribution is symmetric, as the graph reveals. The
standard deviation is 0.53. The analysis of the frequency tables indicates that
62% of preferences are distributed around the mean value.
113
Figure 31: Distribution of raw utilities for FITs
The second preferred attribute level amongst those investigated was the
support scheme “tradable green certificates”, also known as “renewable
portfolio standard”. The distribution of raw utilities is shown in Figure 32. The
frequencies vary from a minimum value of 0.04 to a maximum of 0.96. The
values of the 25° and 75° percentiles are 0.44 and 0.60. The mean, median
and mode of the distribution are, respectively, 0.52, 0.51 and 0.37. The
standard deviation is quite low, with a value of 0.16. In fact, 72% of the raw
utilities are clustered within one standard deviation of the mean.
As far as the asymmetry value and kurtosis are concerned, these are
respectively -0.14 and 1.55. These values indicate a quite symmetric
distribution with a leptokurtic shape.
114
Figure 32: Distribution of raw utilities for RPS, TGCs
Figure 33 illustrates the distribution of raw utilities for tender schemes.
Figure 33: Distribution of raw utilities for tender schemes
115
The distribution shows a minimum value of -1.41 and a maximum value of 0.30. The values of the 25° and 75° percentiles are -1.02 and -0.56. The mean
is -0.77, the median is -0.79, the minimum level of the mode is -1.41.
The standard deviation is 0.25. As the graph shows, only about 57% of the
utilities are concentrated between one standard deviation of the mean, the
remaining values are quite dispersed. By looking at the asymmetry (-0.18) and
kurtosis (-0.68), it is possible to understand that this distribution is slightly
platikurtic.
The least preferred level amongst the support scheme category is
represented by tax incentives and investment grants. As shown in Figure 34,
the raw utilities range from a minimum value of -2.26 to a maximum of -0.98.
The mean and median are respectively -1.78 and -1.79. The values of the 25°
and 75° percentiles are -2.02 and -1.60.
Figure 34: Distribution of raw utilities for tax incentives, investment
grants
The curve has a standard deviation of 0.31. About 65% of utilities are
concentrated within one standard deviation of the mean.
The distribution is characterised by a positive asymmetry (0.47) with a right
tail, and by a platikurtic shape (kurtosis=-0.51).
116
7.2.3
The policy attribute “duration of the support” is articulated in three different
levels: support incentive paid for more than 20 years, 10 to 20 years, and less
than 10 years. Figure 35 shows the results of the distribution for the first level
described. As it is possible to see, the raw utility values range from a minimum
of 1.01 to a maximum of 2.88. The values of the 25° and 75° percentiles are
1.58 and 2.04. The mean is 1.82, the median is 1.80, the minimum level of the
mode is 1.01. The standard deviation is 0.38.
Figure 35: Distribution of raw utilities for more than 20 years
The analysis of frequency reveals that 70% of the sample is clustered within
one standard deviation of the mean value. There is a positive asymmetry
(0.14) and the level of kurtosis is 0.25, indicating a leptokurtic distribution.
Figure 36 illustrates the results for the second level of the duration of incentive
analysed, corresponding to 10-20 years. The distribution is comprised
between the minimum value of -0.39 and the maximum value of 0.43. The
mean and the median have both a value of 0.07. Multiple modes are observed,
with a minimum value of -0.39. The values of the 25° and 75° percentiles are 0.04 and 0.20. The standard deviation is 0.17. The utilities are quite
distributed around the mean value; in fact 73% of frequencies fall within one
standard deviation of the mean, as the graph indicates. There is a negative
asymmetry, with a value of -0.33. Finally, the kurtosis is 0.37.
117
Figure 36: Distribution of raw utilities for 10- 20 years
Figure 37: Distribution of raw utilities for less than 10 years
The third level of this policy attribute was an incentive paid for less than 10
years. Figure 37 reports the results for this distribution.
118
The mean value is -1.89, the median is -1.92 and the minimum mode is -2.79.
Minimum and maximum values for the distribution are, respectively, -2.79 and
-0.92. The standard deviation is 0.38. The analysis of the frequency table
reveals that 68% of utilities are concentrated within one standard deviation of
the mean. The asymmetry is 0.27 and the kurtosis is 0.10.
7.2.4
As far as the length of the administrative process is concerned, three
alternative options were offered to respondents: an iter shorter than 6 months
(“quick”), during from 6 to 12 months (“medium”), or longer than 12 months
(“slow”).
Figure 38 shows the distribution of raw utilities in the first case. The values
range from a minimum of 0.95 to a maximum of 2.39. The mean is 1.57 and
the median is 1.48. The values of the 25° and 75° percentiles are 1.32 and
1.79.
Figure 38: Distribution of raw utilities for less than 6 months
The standard deviation is 0.35. As it is possible to see from the graph, the
frequency values are quite concentrated around the mean. Indeed, 73% of
frequencies fall within one standard deviation.
119
The graph also indicates a positive asymmetry (0.76). The kurtosis is -0.16,
meaning that the values are quite concentrated and not dispersed toward the
tails.
Figure 39 illustrates the distribution for the second level observed. The
minimum level of the distribution is -0.38 and the maximum is 0.39. The mean
has a value of 0.06, while the median is 0.04. Multiple modes are observed
(the minimum value is -0.38). The values of the 25° and 75° percentiles are 0.04 and 1.14.
The standard deviation has a very low value (0.15). About 65% of frequencies
fall within one standard deviation of the mean.
The distribution is quite symmetric, with an asymmetry value of 0.13, and a
kurtosis of 0.13.
Figure 39: Distribution of raw utilities for 6-12 months
Finally, the distribution of raw utilities for the last level of the administrative
process proposed is displayed in Figure 40. This distribution has a minimum
value of -2.65 and a maximum of -0.85. The 25° and 75° percentiles are
respectively -1.85 and -1.37. The mean is -1.63, the median -1.56 and the
minimum level of the mode is -2.65. The standard deviation is 0.40. About
68% of frequencies fall within one standard deviation of the mean. The
asymmetry value is -0.62, and the kurtosis is -0.10.
120
Figure 40: Distribution of raw utilities for more than 12 months
7.2.5
Social acceptance
The last policy attribute which was investigated during the survey is social
acceptance. Two levels are given: high social acceptance, characterized by
pro-wind activism of NGOs, favourable press, pro-wind citizens’ coalitions,
and low social acceptance, where anti-wind activism, negative press and antiwind demonstrations are in place.
The distribution of raw utilities in the first case is shown in Figure 41. The
values range from a minimum of 0.28 to a maximum of 2.35. The 25° and 75°
percentiles are respectively -0.95 and 1.64. The mean of the distribution is
1.27, and the median is 1.25. Also in this case, multiple modes are observed
(the minimum value is 0.28). The analysis of the frequency table indicates that
68% of utilities are concentrated around one standard deviation of the mean
(the standard deviation is 0.44). The asymmetry value is 0.17, and the
kurtosis is -0.23.
Figure 42 shows the results for the low social acceptance case. The values
range from a minimum of -2.35 to a maximum of -0.28. The 25° and 75°
percentiles are respectively -1.64 and -0.95. The mean of the distribution is 1.27, and the median is -1.25. Also in this case, multiple modes are observed
(the minimum value is -2.35). The analysis of the frequency table indicates
that 68% of utilities are concentrated around one standard deviation of the
121
mean (the standard deviation is 0.44). The asymmetry value is -0.17, and the
kurtosis is -0.23.
Figure 41: Distribution of raw utilities for high social acceptance
Figure 42: Distribution of raw utilities for low social acceptance
122
7.3
Sample segmentation
In order to better capture the differences in the share of preferences for
renewable energy policy attributes amongst the various profiles of investors, a
series of analyses through sample segmentation have been conducted using
the Market Simulator tool of the Sawtooth software package. The survey
responses have been elaborated according to the following criteria: type of
companies, experience in the renewable energy investing domain, share of
renewable energy technologies in the portfolio, portfolio composition, working
experience of respondents, and academic background of respondents.
Results are presented and discussed in the following paragraphs.
7.3.1
Influence of company type
To assess the influence of company type, respondents have been regrouped
in the following three categories:
x
Venture Capital, Private Equity or Hybrid combinations (including
private or family-run VC and PE funds)
x
Banks, Hedge Funds, Pension Funds, Personal Funds/Private
Companies and Insurance Companies
x
Other (Project developers and utilities, Infrastructure Funds,
Engineering companies)
It is worth highlighting that the total number sums up to 55, since 5 out of the
60 investors who have completed the ACA section of the questionnaire have
provided no response to the related demographic question and are therefore
excluded from the segmentation analysis.
Results are provided in Figure 43, which displays the average attribute
importances for the sample and the selected sub-samples, and in Table 17,
which reports the comparison of part worth utilities.
123
Figure 43: Average importance of policy attributes for the sample and
for the selected sub-samples
Table 17: Comparison of part worth utilities between the overall sample
and the selected sub-samples
Zero-centered
Sample
VC / PE
total
N=60
N=32
100 €/MWh
60.21
61.64
75 €/MWh
23.60
24.17
50 €/MWh
-17.91
-18.94
25 €/MWh
-65.91
-66.87
RE support scheme
Feed-in tariffs
57.35
57.85
TGC/RPS
14.57
14.68
Tender schemes
-21.85
-21.97
Tax incentives/
-50.06
-50.56
Investment grants
> 20 years
51.21
52.02
10-20 years
2.13
1.88
< 10 years
-53.35
-53.90
<6 months
44.45
44.11
6-12 months
1.49
1.14
>12 months
-45.93
-45.25
Social acceptance
High
35.77
33.91
Low
-35.77
-33.91
Banks
Others
N=10
N=13
63.91
24.75
-18.04
-70.63
56.09
19.76
-16.95
-58.90
56.85
12.79
-21.53
-48.11
58.53
14.03
-23.27
-49.28
49.49
0.19
-49.69
49.82
3.95
-53.77
43.43
2.01
-45.44
43.09
3.24
-46.34
36.22
-36.22
42.09
-42.09
124
As far as the category of venture capital and private equity funds is concerned,
the results of the segmentation analysis reveal that their share of preferences
is quite in line with the average for the whole sample. This indeed is not
surprising since venture capital and private equity funds represent the largest
majority of the represented population in the sample therefore they have a
higher weight compared to other categories of investors.
Looking at the average importances, it is worth noticing that the scale of
preferences is the same for this sub-group and the total sample. The only
difference which is worth pointing out is that venture capital and private equity
funds seem to assign a relatively lower importance to social acceptance
compared to the other investors’ categories.
Banks and the other financial actors included in this sub-group show some
differences compared to the sample average. The level of the premium
incentive is ranked first and receives a higher share compared to the sample
average (27% against 25%). Looking at the part worth utilities, there is a
stronger preference for high levels of the premium incentive (100 €/MWh and
75 €/MWh) compared to the sample average. The type of the renewable
energy support scheme, the duration of the support and the length of the
administrative process are less relevant attributes for this category of
investors compared to the overall sample. Quite remarkably, social
acceptance seems to have a certain influence in the decision to invest in
renewables for this sub-group and is assigned a higher value than the length
of the administrative process, which is the least relevant attribute.
Finally, the category of project developers, utilities, infrastructure funds and
engineering companies displays some interesting peculiarities. For instance,
the level of support seems to be less significant for this sub-group (23%
against 25% for the overall sample), whereas social acceptance receives a
higher ranking (17% against 14%). All remaining attributes seem to be well in
line with the average for the sample.
The analysis of part worth utilities indicates that this sub-group is particularly
attracted by feed-in tariffs compared to other renewable energy support
schemes, and that they would be more willing to accept a shorter duration of
the support (between 10 and 20 years) and a longer length of the
administrative process (between 6 and 12 months) compared to the other
investors.
7.3.2
Influence of the level of exposure to the RE investing domain
As already presented in the descriptive statistics section (chapter 6), two
thirds of investors in the sample have already invested in the renewable
125
energy sector, while one third has not yet invested in this market. Therefore a
segmentation analysis has been carried out in order to see whether there is
any significant difference in the share of preferences between RE and non RE
investors.
As illustrated in Figure 44, there does not seem to be significant differences
between the two categories. Indeed, in both cases the share of preferences
seems to be well aligned with those expressed by the sample average, with
only minor deviations. In particular, non RE investors show less interest for
the duration of support while RE investors assign a greater weight to this
policy attribute compared to the average. As for social acceptance issues, one
possible explanation for this result is that non RE investors tend to
overestimate this aspect compared to RE investors due to lack of knowledge
of the sector, therefore being more influenced by negative press or anti-wind
activism.
Additional qualitative research would be needed in order to provide a careful
explanation for this phenomenon.
Figure 44: Importance of policy attributes for the sample and for RE and
non-RE investors
7.3.3
Influence of the share of renewables in the portfolio
Figure 45 below presents the results of a segmentation analysis where the
sample has been split according to the share of renewable energy
technologies in the investment portfolio. As one can see, the level of the
incentive granted to renewable energy project is the most important attribute,
regardless of the share of renewables composing the portfolio. Also the
126
ranking for the remaining policy attributes remains fairly homogeneous across
the sub-sample, and in line with the sample average. The only slight deviation
that can be observed is that those companies which only invest in renewables
seem to assign lower importance to the length of the administrative process
and to social acceptance compared to the other categories and to the sample
average.
This corroborates the previous finding, i.e. that those investors who have
more familiarity with the renewable energy business tend to be less
concerned by social acceptance issues.
Companies investing only in renewables consider the duration of the support
given to the renewable energy project a very important policy attribute, more
important that the type of policy support scheme implemented per se. It is
worth noticing that no clear correlation can be identified between the ranking
assigned to the various attribute levels and the share of renewables in the
portfolio.
Figure 45: Importance of policy attributes for the sample and for RE
investors
7.3.4
Influence of portfolio composition
In order to gather more analytical insights on what are the policy attributes
which influence the most the likelihood to invest in single renewable energy
technologies, an additional segmentation has been carried out, by identifying
a sub-group of respondents investing exclusively on one renewable energy
technology. It is worth highlighting that the sample is small (only 10
127
respondents have such characteristics), therefore this segmentation analysis
needs to be interpreted with care. Nevertheless it provides useful indications
for future research. In fact, as shown in Figure 46 policy attributes seem to
have different impacts on respondents depending on the type of technology
they are investing in.
Figure 46: Importance of policy attributes for the sample and for
respondents investing only on one specific technology
One remarkable finding is that the length of the administrative process and
social acceptance issues play a much greater role on the decision to invest in
biofuels rather than other renewable energy technologies. These two
attributes are perceived to be much more relevant than other attributes
influencing the profitability of the investment (such as the level of incentive
and the duration of support). As a matter of fact, the current debate around
the sustainability of biofuels has a direct impact on the framework conditions
to invest in this technology; this can provide an explanation for such peculiar
scales of preferences. Further research is recommended in order to see if
these preliminary findings are validated or not based on a larger sample of
investors operating in the biofuel market. Respondents investing either on
solar PV or on wind technologies assign a higher preference to the level of
support, type of support scheme and duration of the support compared to the
sample average, and a lower preference to the administrative framework and
social acceptance, thus confirming earlier results.
128
7.3.5
Influence of experience
The influence of experience on investors’ preferences for different renewable
energy policies is assessed in two ways: i) by looking at the number of years
of experience in the renewable energy investing domain, and ii) by looking at
the investors’ age as a proxy for their overall experience in the investment
sector.
As far as the first aspect is concerned, Figure 47 reveals that there is no
significant relation between preferences for specific policy attributes and
experience. The only remarkable exception is related to the administrative
framework; indeed, the importance of this policy attribute decreases with the
increasing of the number of years of experience in the renewable energy
market. It is worth noticing that the most experienced investors (more than 10
years) assign a higher value to the type of renewable energy support scheme
and to the duration of the support compared to the others and to the sample
average.
Figure 47: Importance of policy attributes for the sample and for the subsample split by years of experience in the RE investing domain
By looking at Figure 48, it can be seen that there is no relationship between
preferences for specific policy attributes and the age of respondents. This
seems to support the finding that investors’ preferences for different policy
types are influenced by the specific experience they have gathered in the
renewable energy industry business, rather than by their overall working
experience.
129
Figure 48: Importance of policy attributes for the sample and for the subsample split by age of respondents
7.3.6
Influence of the background
The analysis of investors’ academic background offers interesting insights.
Compared to the sample average, the type of renewable energy support
scheme receives a high ranking by all categories with the only exception of
legal experts. The latter assign a much higher importance to social
acceptance issues and to the level of the incentive compared to the other
respondents.
Level and duration of the support provided to renewables seem to have the
same importance for those respondents who have a background either in
economics and business administration or in engineering.
Results are shown in Figure 49.
130
Figure 49: Importance of policy attributes for the sample and for the subsample split by academic background
7.4
Renewable energy optimism and preference for policy
schemes
In the present paragraph, some complementary findings on the relationship
between the investors’ beliefs about the technological and market potential of
renewable energies, and their preferences for renewable energy policy
schemes are presented, based on additional statistical analysis of the sample
responses.
The degree of “renewable energy optimism” was assessed by looking at the
responses given by the sample to the questions regarding some technological
and market beliefs, as reported in section two of the questionnaire. In
particular, a “renewable energy optimism” index was built by factor-analysing
the variables related to a high penetration rate of renewables (up to 80% and
100%) in Europe by 2050, the variable related to the growth potential of
renewables worldwide in the next 20 years, and the variable related to the
competitiveness of solar PV electricity within the next decade.
The sample was split in two categories: “RE optimists”, characterised by a
value higher than the mean, and “RE pessimists”, if the ranking was equal or
below the mean.
The degree of renewable energy optimism was measured against the utility
values assigned by investors to the policy attribute “type of RE support
scheme”, as investigated in the ACA section of the questionnaire. To this end,
131
the sample was split in two categories: the first regroups those respondents
whose individual utilities assigned to the various policy attributes are above
the mean, whereas the second is composed of respondents whose individual
utilities are equal or below the average for the sample.
Figure 50 shows the distribution of utilities for the attribute “type of RE support
scheme”. About 17% of the investors fall in the category characterised by a
low degree of RE optimism and low degree of confidence in RE support
mechanisms. Conversely, an additional 38% of “RE pessimists” have a high
degree of confidence in RE support mechanisms, thus implying that only a
strong policy intervention can help overcome technical and market barriers for
renewables. About 15% of the sample has a high degree of RE optimism but
a low degree of confidence in RE support schemes, thus displaying a certain
policy aversion, whereas 30% of investors have both a high degree of RE
optimism and a high degree of confidence in RE support mechanisms.
(N=60)
Figure 50: RE optimism and confidence in RE support schemes
Figure 51 and Figure 52 look more in detail at the two most important types of
policies for the sample, i.e. feed-in tariffs and tradable green certificates.
Following the procedure already described above, the degree of confidence in
feed-in tariffs and in tradable green certificates has been measured by looking
at the individual utilities, as revealed by the ACA analysis. A high degree of
confidence in feed-in tariffs (or tradable green certificates) corresponds to
individual utilities above the mean, whereas a low degree of confidence
corresponds to utility values below or equal to the mean.
It can be seen that the sample is fairly homogeneously distributed. In the case
of feed-in tariffs, in fact, about one fourth of the investors show both a RE
132
pessimism and a low degree of confidence in this type of RE policy scheme;
32% of the sample is characterised by a low degree of RE optimism but a high
degree of confidence in feed-in tariffs, 21% have a high degree of RE
optimism but a low confidence in FITs, and 23% are both RE optimists and
have a high confidence in FITs.
(N=60)
Figure 51: RE optimism and confidence in feed-in tariffs
(N=60)
Figure 52: RE optimism and confidence in tradable green certificates
In the case of tradable green certificates, the percentage of pessimists is
slightly higher: in fact 32% of the sample display a low degree of RE optimism
and a low degree of confidence in TGCs; 25% have a high degree of
confidence in TGCs while being RE pessimists, 17% are RE optimists and
have a low degree of confidence in TGCs and 26% are RE optimists and have
a high degree of confidence in TGCs.
133
7.5
Degree of confidence in market efficiency and
preference for policy schemes
The degree of confidence in market efficiency was measured by looking at the
ratings given by the investors to the following two statements:
x
Government intervention does more harm than good, let
governments stay out of the way (used as a proxy for a high degree
of confidence in market efficiency), and
x
Market forces alone will never lead to a significant exploitation of
renewables (used as a measure for low confidence in market
efficiency)
The purpose was to understand whether investors displaying a higher
confidence in the role of markets have a stronger preference for a more
market-based RE support scheme like green certificates compared to feed-in
tariffs, and if, conversely, less market confident investors (ergo investors
displaying a more pro-government attitude) assign higher utility values to
feed-in tariffs.
The analysis reveals that indeed there is a correlation between a certain
dislike of government intervention and a stronger preference for more marketoriented renewable energy support schemes. For example, by looking at
Figure 53 it can be seen that those investors who agree or strongly agree with
the statement reported, also have a stronger preference for tradable green
certificates.
On the other end, those investors who do not agree with the same statement
have stronger preference for feed-in tariffs, as shown in Figure 54.
As far as the lower degree of confidence in market efficiency (or more progovernment attitude) is concerned, the link with a higher preference for feed-in
tariffs is clear (see Figure 55), whereas no significant relationship can be
identified with the preference for tradable green certificates, as displayed in
Figure 56.
134
Figure 53: High confidence in market efficiency and preferences for
TGCs
Figure 54: High confidence in market efficiency and preferences for FITs
135
Figure 55: Low confidence in market efficiency and preferences for FITs
Figure 56: Low confidence in market efficiency and preferences for
TGCs
136
In order to get a better understanding on the relationship between the degree
of confidence in market efficiency and the preferences for more market or
government-based support schemes, the information has been further
elaborated and summarised in the following matrices. The index of confidence
in market efficiency was built by looking at the ratings assigned to the two
statements reported above, while the preference for TGCs was calculated
according to the utility values assigned to this policy attribute (below, equal or
above the mean).
According to Figure 57, 43% of investors have a low degree of confidence in
market efficiency and assign low utilities to TGCs. This might seem to confirm
that more pro-government investors have a lower preference for marketbased instruments. However, another 39% of investors have a low degree of
confidence in market efficiency but a high level of confidence in TGCs. At the
other extreme, 14% of investors have a high degree of confidence in market
efficiency and have a high preference for TGCs, while 4% have a high degree
of confidence in market efficiency and have a low preference for TGCs.
Figure 57: Distribution of the sample according to the degree of
confidence in market efficiency and preferences for TGCs
Figure 58 reveals that those investors who have a low degree of confidence in
market efficiency also assign high utility values to FITs (64% of the sample).
This confirms the previous finding, i.e. that pro-government investors look for
more regulation in the renewable energy market. Another 25% of investors
have low confidence in market efficiency but a low level of confidence in FITs.
At the other extreme, 11% of investors have a high degree of confidence in
market efficiency and have a low preference for FITs, while none of the
investors having a high degree of confidence in market efficiency has also a
high degree of confidence in FITs.
137
Figure 58: Distribution of the sample according to the degree of progovernment attitude and preferences for FITs
7.6
Assessing the likelihood to invest in different policy
scenarios
The market simulator tool allows to investigate how the likelihood to invest in a
renewable energy project might change depending on the variation of each
attribute level. To this end, a base case scenario has been defined, based on
the utility values assigned by respondents and already presented in paragraph
7.1. The ACA analysis in fact has shown that the level of the support is the
most important attribute, followed by the type of renewable energy support
scheme, the duration of the support, the length of the administrative process
and social acceptance. Additionally, it has revealed that among the proposed
renewable energy support schemes, feed-in tariffs are by far the most
preferred mechanisms. Based on these elements, a base case policy scenario
has been selected, which is characterized by the following attribute levels:
x
Type of policy scheme: feed-in tariff
x
A premium incentive of €/MWh 50
x
A duration of the support longer than 20 years
x
A length of the administrative process of 6-12 months
x
A high social acceptance
As far as the type of policy scheme is concerned, feed-in tariff has been
chosen in the base case scenario both because it is the most relevant support
scheme for the surveyed investors and because it is the most widespread
policy instrument in Europe. Regarding the premium incentive, the average
tariffs for onshore wind paid in EU countries range between 50 €/MWh and 75
138
€/MWh, with Italy being closer to the highest attribute level proposed (100
€/MWh) and Spain and Denmark being closer to the lowest attribute level
listed (25 €/MWh). Therefore, in the base case a premium incentive of 50
€/MWh has been selected, which reflects the average premium incentive paid
for onshore wind in a country like Germany over the past years. The base
case scenario also assumes that a stable framework is in place, offering the
support for more than 20 years, and a reasonable length of the administrative
process, in line with the findings of the OPTRES study (OPTRES, 2006)
already commented in paragraph 7.1. Finally, by default a high social
acceptance of wind energy is assumed, again in line with the results of the
OPTRES survey. Once defined the attributes of the reference scenario, a
series of simulations have been run by changing one more parameters at a
time. The results are summarized in Table 18.
characterized by a very high level of the incentive (100 €/MWh) and
varying levels of other attributes
Simulation
Scenario
Base case
st
1 simulation
nd
2 simulation
rd
3 simulation
th
4 simulation
th
5 simulation
Likelihood to invest
6%
Base case + 100 €/MWh
incentive
94%
Base case
27%
incentive and 10-20 yrs support
duration
73%
Base case
70%
incentive and <10 yrs support
duration
30%
Base case
26%
incentive and length of the
admin process >12 months
74%
Base case
45%
incentive and low social
acceptance
55%
Since the level of the premium incentive turned out to be the most important
policy attribute, in the first simulation the base case scenario is compared
139
against an alternative scenario characterized by a very high level of incentive
(100 €/MWh) and varying levels of the other attributes. As reported in Table
18 the second scenario is much more attractive than the reference case,
given the higher remuneration provided to the renewable energy project,
obtaining 94% of preferences all other attributes being equal. However, by
changing the level of the remaining attributes, the alternative scenario
becomes the less and less attractive. For instance, if the duration of the
support guaranteed to the renewable energy project is shortened from more
than 20 years to 10-20 years, the share of preferences for the alternative
scenario is of 73%, despite the very high level of the incentive. Quite
remarkably, if the duration of the support is shortened to less than 10 years,
the base case scenario becomes much more preferable than the alternative
one, even if the amount of the incentive provided to the renewable energy
project is exactly the half. This supports the results provided above, and
represents a very important indication for policy design. Indeed, it seems that
a timeframe of 10-20 years is the minimum to guarantee the stability of the
market and to attract investments in the sector. Stop and go policies, or
frequent changes in the policy design are very detrimental to the growth of the
renewable energy market, since they create uncertainty and discourage
investments.
As for the length of the administrative process, the simulation confirms what
has been presented already in the previous sections, i.e. that the
administrative framework is less of a concern for investors than other policy
attributes. In fact, even if the duration of the administrative process becomes
longer than one year, still the alternative scenario is more preferable than the
base case. This finding suggests that respondents are ready to spend a
certain amount of time after the bureaucratic and authorization procedures,
provided that they receive enough remuneration for their investment.
Finally, in the case social acceptance becomes low, the base case and the
alternative scenario receive almost the same share of preferences. This is an
indirect confirmation of previous findings. It seems in fact that in general
investors (and particularly the most experienced ones) believe that social
acceptance is not an issue for onshore wind in Europe. However, in a
scenario characterised by strong opposition, social acceptance becomes an
important attribute to be considered while taking an investment decision.
As reported in Table 19, if the incentive premium is lowered to 75 €/MWh, the
base case scenario becomes preferable over the alternative scenario already
if the duration of the support is shortened to 10-20 years (57% against 43%).
140
In the case the support duration is shorter than 10 years, the base case
scenario receives 90% of preferences.
The duration of the administrative process has a much higher impact in the
case the amount of the incentive level of 75 €/MWh. In this case in fact the
base case scenario is preferred over the alternative scenario if the latter is
characterized by a longer duration of the administrative process. In other
terms, investors prefer obtaining 25 €/MWh less provided that the
administrative process is ended within one year.
Finally, the base case is definitively more attractive than the alternative
scenario if social acceptance in the country is low (71% against 29%).
characterized by a high level of the incentive (75 €/MWh) and varying
levels of other attributes
Simulation
st
1 simulation
nd
2 simulation
rd
3 simulation
th
4 simulation
th
5 simulation
Scenario
Base case
incentive
Base case
incentive and 10-20 yrs support
duration
Base case
incentive and <10 yrs support
duration
Base case
admin process >12 months
Base case
incentive and low social
acceptance
19%
81%
57%
43%
90%
10%
54%
46%
71%
29%
Two additional simulations are offered in Table 20, which compares the base
case against a scenario characterized by a very low level of the premium
incentive (€ 25/MWh). Quite obviously, the base case receives 83% of
preferences compared to the alternative scenario, given the more attractive
level of incentive. However, if the low level of the incentive in the alternative
141
scenario is compensated by a quick administrative process, shorter than 6
months, the two scenarios become comparable.
This finding confirms that investors are primarily driven by financial
considerations on the profitability of their investment, but some of them are
ready to accept lower remunerations if a more rapid administrative process is
guaranteed.
characterized by a low level of the incentive (25 €/MWh) and a quick
Simulation
st
1 simulation
nd
2 simulation
7.7
Scenario
Base case
incentive
Base case
admin process <6 months
83%
17%
54%
46%
Sensitivity analysis: the influence of the estimate
method
The Hierarchical Bayes is recommended in the analysis of part worths since it
proves to produce more robust results compared to traditional estimation
methods, like the ordinary least squares. As highlighted by several scholars
(e.g.: Allen, 2004; Lenk et al.,1996), when the goal is to estimate the
heterogeneity in the customers' part-worths, the least squares method
requires each subject to respond to a higher number of product profiles than
product attributes, resulting in lengthy questionnaires and in higher risk of
biased responses. Conversely, the Hierarchical Bayes models do not require
the individual-level design matrices to be of full rank. This leads to the
possibility of using fewer profiles therefore having shorter questionnaires. The
Hierarchical Bayes method therefore helps in situations where the data
collection task is so large that the respondent cannot reasonably provide
preference evaluations for all attribute levels. In fact the Hierarchical Bayes
approach uses average information about the distribution of utilities from all
respondents to estimate attribute level utilities for each individual. This
approach again allows more attributes and levels to be estimated with smaller
amounts of data collected from each individual respondent. Having said that,
for sake of completeness the main results for the sample calculated with the
142
ordinary least squares method are reported as well. This allows to assess the
influence of the estimate method on the results.
Table 21: Average importance of policy attributes for the investigated
sample, according to the estimate method selected
Policy
attribute
Level of the
support
RE support
scheme
Duration of the
support
Length of the
administrative
process
Social
acceptance
HB
OLS
Average
importance
Standard
deviation
Average
importance
Standard
deviation
25.22%
4.32
26.58%
8.68
21.48%
5.07
21.07%
7.23
20.91%
3.98
19.43%
9.17
18.08%
4.38
16.69%
8.65
14.31%
5.15
16.23%
8.10
100.00%
100.00%
As it is possible to see from Table 21, the ranking order of the average
importance remains the same in the two cases. However, the attribute
importance weight is different. In particular, via the Ordinary Least Square
method the level of the support receives an even higher score than through
the Hierarchical Bayes. Conversely, the importance assigned to the
administrative process is much lower. It is worth signalling that social
acceptance receives a higher weight by using the OLS method. Although it
does not impact the overall ranking order, this is a very significant difference
(16.23% against 14.31%) which makes the importance of social acceptance
become very close to the length of the administrative process if using the OLS
instead of the HB method. This should be kept in mind while selecting
different elaboration techniques. Sensitivity analysis is therefore always
commendable. Finally, the standard deviations are higher in the OLS results
than in the HB, thus confirming the analytical superiority of the latter tool.
Looking at the part-worth utilities (Table 22), it can be noted that some
differences are less pronounced while using the OLS method, while others
increase even further. For example, the distance between FITs and TGCs
increases. Nevertheless, the absolute distance between the best and worst
level is shorter compared to the HB analysis results. In terms of the duration
of support and duration of the administrative process, the distance between
143
the offered levels is shorter than in the HB results. Finally, as already
discussed above, part-worth utilities for social acceptance are higher when
using OLS compared to HB.
Again, it is worth highlighting that the standard deviations are higher in the
OLS results than in the HB estimates.
Table 22: Importance of policy attribute levels for the investigated
sample, and standard deviations calculated with HB and OLS method
ZeroHB
centered
Av. utilities
Sd
100 €/MWh
60.21
10.11
75 €/MWh
23.60
7.33
50 €/MWh
-17.91
4.17
25 €/MWh
-65.91
15.67
RE support scheme
Feed-in tariffs
57.35
13.98
TGC/RPS
14.57
4.72
Tender
-21.85
6.63
schemes
Tax
incentives/
-50.06
8.22
Investment
grants
More than 20
51.21
10.20
years
10-20 years
2.13
5.10
Less than 10
-53.35
10.32
years
<6 months
44.45
10.73
6-12 months
1.49
4.36
>12 months
-45.93
11.58
Social acceptance
High
35.77
12.88
Low
-35.77
-12.88
OLS
Av. utilities
Sd
58.25
24.44
-17.61
-65.07
28.61
24.66
25.28
28.61
52.78
5.74
29.75
27.56
-24.47
24.16
-34.04
27.61
43.43
28.14
4.11
20.24
-47.54
28.71
37.27
2.42
-39.68
27.06
18.65
24.90
39.71
-39.71
21.92
21.92
144
8
Results of the regression model
The previous chapters have provided descriptive statistics for the sample, as
well as an analysis of policy preferences.
In this section the conceptual model is tested against the data collected in
order to see if and to what extent the hypotheses are confirmed, or if they
need to be rejected. To this end, the two equations described in paragraph
5.3.2.3 (page 74) have been tested using several estimation methods.
8.1
Most relevant findings for the sample
First and foremost, it is worth signalling that the empirical study depicts a
business context dominated by highly rational and well informed investors,
who tend to minimise the risk of their choices by founding their decisions on
factual, technical information, as already reported in chapter 6. The analysis of
the survey results carried out in the previous chapters has revealed that the
investors in the sample have clear preferences for relatively mature renewable
energy technologies (as displayed in Table 6). They also look for stable policy
frameworks axed on high levels of financial support as a prerequisite for
investment decisions.
Against this background, the multivariate regression analysis has brought
some additional findings to light. In particular, it has identified a series of
causal relationships between behavioural attitudes and the share of
renewables in the investment portfolio, as well as between behavioural
attitudes and the investment performance, thus addressing the two research
questions.
Table 23 displays the descriptive statistics for the investigated variables and
Pearson’s correlations, while the results of the multivariate regressions for the
renewable energy share model and the investment performance model are
reported in Table 24 and Table 25, respectively. It is worth remembering that
in order to exclude possible endogeneity problems in the data structure, the
model was tested using four different estimation methods: the Ordinary Least
Squares (OLS), 2-stage Least Square (2SLS), 3-stage Least Square (3SLS),
13
and Seemingly Unrelated Regression (SUR) as well .
For sake of simplicity in the following discussion I refer to the OLS results only.
As can be seen from the Tables, the estimates are consistent across the
13
Please refer to chapter 5 for a thorough explanation of this methodological issue.
145
different methods applied. This confirms the robustness of the conceptual
model developed.
In order to take duly into account the methodological limitations deriving from
the use of a qualitative measurement method based on a three-point likert
scale in the assessment of performance, the equation 2 reported at page 74)
was re-estimated using also a multinomial logit model. In fact, as the
Hausman test revealed no endogeneity problems, it was possible to estimate
equation 1 and 2 separately.
The estimates of the logit model are consistent with the original OLS
estimates, thus confirming the validity of results. For sake of completeness,
they are reported in Table 26, but not commented in the interest of space.
Eventually, additional analysis based on quantitative longitudinal data sets will
be needed in order to further corroborate the findings.
146
Table 23: Descriptive statistics and Pearson Correlations
Confidence in
market
efficiency
Confidence in
technology
adequacy
Technological
risk seeking
attitude
Perceived
importance of
policy type
Perceived
importance of
support level
Perceived
importance of
support duration
Perceived
importance of
the length of the
administrative
process
Investor’s
experience
RE share in the
investment
portfolio
Investment
performance
Min
Max
Mean
Std
Confidence
in market
efficiency
Confidence
in
technology
adequacy
Technological
risk seeking
attitude
Perceived
importance
of policy
type
Perceived
importance
of support
level
Perceived
importance
of support
duration
Perceived
importance of
the length of
administrative
process
Investor’s
experience
RE share
in the
investment
portfolio
1.00
5.00
3.45
0.79
1
-0.03
0.03
0.17
0.21
0.20
0.15
0.22
0.37
2.00
5.00
3.56
0.72
-0.03
1
0.06
0.08
-0.02
-0.06
0.02
0.12
0.13
0.00
9.00
0.48
1.02
0.03
0.06
1
0.03
-0.08
-0.05
-0.01
-0.17
-0.15
0.71
5.17
1.68
0.78
0.17
0.08
0.03
1
0.85
0.88
0.81
0.06
0.14
0.64
8.86
2.02
1.19
0.21
-0.02
-0.08
0.85
1
0.94
0.82
-0.01
0.04
0.48
6.91
1.67
0.95
0.20
-0.06
-0.05
0.88
0.94
1
0.81
0.08
0.15
0.49
4.34
1.35
0.56
0.15
0.02
-0.01
0.81
0.82
0.81
1
-0.05
0.07
1.00
4.00
2.29
0.62
0.22
0.12
-0.17
0.06
-0.01
0.08
-0.05
1
0.34
0.00
5.00
2.11
1.94
0.37
0.13
-0.15
0.14
0.04
0.15
0.07
0.34
1
1.00
3.00
2.24
0.41
0.21
-0.11
-0.32
0.01
0.10
0.13
-0.01
0.08
0.26
147
Table 24: Results of the multivariate regression for the first part of the conceptual model (all coefficients are
standardized)
RE share in the
investment
portfolio
Estimate
OLS
2SLS
SUR
Std
Err
p
Std
Err
p
Std
Err
p
Estimate
Estimate
3SLS
Estimate
Std
Err
p
Confidence in
market
efficiency
0.34
0.24
0.17
0.34
0.24
0.17
0.33
0.24
0.18
0.27
0.25
0.27
Confidence in
technology
adequacy
0.81
0.22
0.00
0.81
0.22
0.00
0.83
0.22
0.00
0.87
0.22
0.00
Technological
risk seeking
attitude
-0.37
0.09
<.0001
-0.37
0.09
<.0001
-0.37
0.09
<.0001
-0.36
0.09
0.00
Perceived
importance of
policy type
0.32
0.43
0.46
0.32
0.43
0.46
0.31
0.43
0.46
0.28
0.44
0.52
Perceived
importance of
support level
-1.34
0.53
0.01
-1.34
0.53
0.01
-1.32
0.53
0.02
-1.19
0.54
0.03
Perceived
importance of
support
duration
1.29
0.59
0.03
1.29
0.59
0.03
1.30
0.59
0.03
1.30
0.61
0.03
Perceived
importance of
the length of
the
administrative
process
0.33
0.58
0.58
0.33
0.58
0.58
0.28
0.59
0.64
0.05
0.60
0.93
148
Table 24: Results of the multivariate regression for the first part of the conceptual model (all coefficients are
standardized) (cont.)
RE share in the
investment
portfolio
Investor’s
experience
Dummy Funds
Dummy VC
Dummy other
investors
R
F
2
OLS
2SLS
SUR
Std
Err
p
Std
Err
p
Std
Err
p
0.52
0.28
0.07
0.52
0.28
0.07
0.51
0.28
0.07
-0.84
0.93
0.37
-0.84
0.93
0.37
-0.86
0.93
0.24
0.60
0.69
0.24
0.60
0.69
0.22
0.60
-0.40
0.59
0.50
-0.40
0.59
0.50
-0.42
0.59
Estimate
0.35
4.14
Estimate
0.35
Estimate
0.35
3SLS
Std
Err
p
0.51
0.28
0.07
0.36
-0.95
0.95
0.32
0.72
0.14
0.60
0.81
0.48
-0.50
0.60
0.41
Estimate
0.35
<0.01
149
Table 25: Results of the multivariate regression for the second part of the conceptual model (all coefficients are
standardized)
Investment
performance
RE share in the
investment
portfolio
Estimate
OLS
2SLS
SUR
Std
Err
Std
Err
Std
Err
p
Estimate
p
Estimate
3SLS
p
Estimate
Std
Err
p
0.05
0.02
0.04
0.09
0.06
0.15
0.06
0.02
0.02
0.09
0.06
0.15
-0.12
0.03
0.00
-0.11
0.03
0.00
-0.12
0.03
0.00
-0.11
0.03
0.00
Investor’s
experience
0.01
0.10
0.89
-0.03
0.11
0.79
0.00
0.10
0.95
-0.03
0.11
0.79
Dummy Funds
-1.12
0.16
<.0001
-1.08
0.18
<.0001
-1.11
0.16
<.0001
-1.08
0.18
<.0001
Dummy VC
-0.93
0.16
<.0001
-0.96
0.17
<.0001
-0.94
0.16
<.0001
-0.96
0.17
<.0001
Dummy other
investors
-0.86
0.13
<.0001
-0.83
0.14
<.0001
-0.85
0.13
<.0001
-0.83
0.14
<.0001
R2
0.23
F
3.61
Technological
risk seeking
attitude
0.20
0.23
0.20
<0.01
150
Table 26: Analysis of the investment performance: results of the logit model
Dependent Variable: Investment Performance
2
Estimate
Std Err
p!&
RE share in the investment portfolio
0.86
0.36
0.02
Technological risk seeking attitude
-0.76
0.43
0.08
Investor’s experience
-0.05
0.33
0.89
Dummy Funds
-14.17
189.40
0.94
Dummy VC
-12.98
189.40
0.95
Dummy other investors
-13.39
189.40
0.94
-2LL
83.60
2
21.33
&
<0.01
151
Both the renewable energy share model and the investment performance
model are significant (F = 4.14 with p < 0.01 and F = 3.61, p < 0.01
2
respectively) and have an acceptable explanatory power (R = 0.35 and 0.23
respectively).
8.2
Main results of the first part of the regression model
Starting from the analysis of Table 24, it can be noted that a priori beliefs have
a positive influence on the investors’ willingness to back renewable energy
projects. However, the degree of confidence in technology adequacy has a
stronger impact than the confidence in the market efficiency (β=0.81 with p <
0.01 versus β=0.34 with p = 0.17).
This might be interpreted as an indication that the proven reliability of a
technology is a conditio sine qua non for investing, whereas investors believe
that possible market inefficiencies can be corrected through the adoption of
appropriate policy instruments. In other terms, investors seem to have a
strong preference for those technologies which have already overcome both
the technology and cash flow “valleys of death” (Grubb, 2004; Murphy and
Edwards, 2003). This finding is in line with the results of a survey carried out
by Fritz-Morgenthal et al. (2009) where the majority of the sample stated that
investors are expected to focus less on innovation and more on established
technologies in the next 2-3 years, as a response to the financial crisis.
Quite surprisingly, and in sharp contrast with the hypothesized effect, a higher
propensity for technological risk (i.e. a tendency to invest in radically new
technologies which are still far from commercial viability) is negatively
associated with the renewable energy share in the portfolio. This can be due
to the fact that most of the portfolios are skewed towards relatively well known
renewable energy technologies. Investors who have an appetite for
technological risk and invest in radically new technologies need to hedge
against this risk by including a higher share of conventional technologies in
their portfolios compared to those who invest in less innovative technologies.
Thus, the total renewable energy share in their portfolios will be lower than
investors with a moderate appetite for technological risk. Another explanation
could be that investors willing to invest in radically new technologies still do
not find enough credible and well documented investment opportunities in the
renewable energy sector in Europe.
Policy preferences have different impacts on the share of renewables in the
investment portfolio. The analysis of their specific effects offers interesting
insights. The influence of the type of policy scheme is statistically not
significant (p = 0.46). As the analysis of the conjoint measurement results has
152
highlighted already, this is due to the fact that the surveyed investors have a
clear and strong preference for feed-in tariffs over other policy instruments. In
other words, the investors in the sample believe that feed-in tariffs are by far
the most effective policy instruments to attract investments in renewable
energy technologies. This limits the variance of the policy type variable and its
consequent effect on investment choices.
A high perceived importance of the level of support is negatively associated
with the share of renewable energy in the investment portfolio (β = -1.34 with
p = 0.01), suggesting that investors implicitly believe that the level of support
currently allocated to renewable energy technologies is still inadequate. In
other words, this means that investors are reluctant to embark in renewable
energy investments and tend to wait until higher levels of support are given.
Conversely, and in sharp contrast with the above result, a high perceived
importance of the duration of the support is positively associated with the
renewable energy share (β = 1.29 with p = 0.03). This suggests that investors
believe that the time horizon of the policies currently in place is already
adequate, and therefore no additional action should be put in place by
governments with respect to this particular policy attribute.
These results seem to reinforce the impression that investors in the sample
are extremely risk averse. The statistical analysis in fact suggests that short
term policies that provide high levels of financial incentives for a limited
amount of time are preferred over long term policies that guarantee a
moderate but stable level of support for a longer amount of time. This finding
should be interpreted with care, because the sample is skewed toward
venture capital and private equity funds, which have rather short-term
investment horizons. The segmentation analysis reported in the previous
chapter has revealed indeed that, for some particular categories of investors
such as infrastructure funds and project developers, the incentive level has a
relatively lower importance, while other policy attributes such as the type of
support scheme and the duration of support play a more important role in
shaping the investors’ preferences.
Finally, the influence of the perceived importance of the length of the
administrative process is statistically not significant (p=0.58).
As far as the control variables are concerned, the investor’s experience has a
slight positive influence on the share of renewables in the investment portfolio
(β=0.52 with p = 0.07), thus confirming that the investors’ confidence in
renewable energy technologies increases with an increase of knowledge in
this specific business. This implies also that those fund managers who have a
153
larger experience in the renewable energy investing domain are more capable
to recognize the value of more innovative technologies, thus being more
inclined to prefer renewables over traditional energy sources.
These results reinforce what the ACA analysis had already suggested, thus
increasing the validity of the empirical findings.
8.3
Main results of the second part of the model
The analysis of the performance model in Table 25 reveals that higher shares
of renewable energy technologies in the investment portfolio are associated
with a slightly higher performance relative to direct competitors (β = 0.05 with
p = 0.04).
This provides evidence against the belief that investments in renewable
energy technologies yield lower returns compared to investments in
conventional energy systems.
This finding appears very relevant for both practice and theory. Firstly,
demonstrating that renewable energy technologies can represent a profitable
investment can help financiers to allocate more capital in this sector, thus
fostering the growth of the renewable energy market. Secondly, this result is
in line with previous findings in management literature, which have highlighted
the positive relationship between sustainability investments and financial
performance. Finally, the result is coherent with portfolio theory, which has
highlighted the positive role of renewable energy investments in risk hedging
strategies. In this respect, the present finding contributes to highlight the
positive role of renewables under the framework of investment diversification
strategies.
As already pointed out, the present results are based on a self-assessed
performance of investors compared to their direct competitors. Even though
the performance equation does not result to be affected by self-reported
biases according to the results of the econometric tests performed, additional
analysis with external data series is recommended to validate the findings in
future research.
It is also interesting to note that the investors’ attitude towards renewable
energy technological risk has a strong negative impact on the investment
performance. This effect is both direct and indirect through its impact on the
renewable energy share in the investment portfolio. Once again, this
reinforces the impression that the surveyed investors display aversion not only
for financial risk, but also for technological risk. One possible explanation for
this result is related to the fact that the majority of respondents in the sample
154
have rather limited experience in the renewable energy sector. Since
investors have not accumulated enough experience in an industry that is very
promising, but also risky, they might fail to analyse investment opportunities in
a proper way, thus privileging short term returns instead of embarking in
projects that guarantee superior returns only in the long run.
As far as the indirect effect of the investors’ attitude towards renewable
energy technology risk is concerned, a possible explanation is that risk
adverse investors tend to include a lower share of renewable energy
technologies in their portfolio, and this leads to a lower performance.
Finally, the type of companies undertaking the investment seems to have a
negative impact on the investment performance. In particular, the results
suggest that - among the three categories of investors - banks, pension funds
and insurance companies are the least performing (β = -1.12 with p < 0.01),
whereas project developers, utilities, infrastructure funds and engineering
companies show better performance.
This makes sense intuitively. In fact, the category of project developers
regroups investors with a specific know how in the renewable energy
investment domain, while banks, pension funds and insurance companies on
average have the least specific knowledge among the three categories
analysed. This finding is in line with the considerations already expressed
above regarding the influence of specific experience on the investment
performance.
155
9
Conclusions
Renewable energy sources can play a crucial role in reducing carbon
emissions and fossil fuel consumption, thus enabling a real energy technology
revolution. Economies of scale, technology improvements and renewable
energy targets are all factors that have contributed to the acceleration of
renewable energy penetration, particularly over the last decade.
Evidence suggests also that renewable energy technologies have long-term
growth prospects and can constitute safe investment opportunities to hedge
against fluctuations of fossil fuel prices.
Despite this promising outlook, their potential is far from being significantly
exploited. Mainstreaming renewable energy technologies requires to inject in
the market an amount of money two to four times higher than the current
values by 2020. Mobilising private capital becomes therefore a priority. Private
investors represent already the largest source of capital for renewable energy
projects. However, they are still reluctant to scale up the share of their
investments in this sector. Needless to say, this is particularly challenging in a
context of global economic slowdown such as the one the world is currently
experiencing. To bridge the gap between the current level of investments and
the actual requirements, significant improvements are needed in terms of
innovative policy design, the development of more accurate evaluation tools
and the understanding of investors’ acceptance of renewable energy.
The literature analysis has emphasized some important aspects. Firstly, it has
highlighted that there is a positive relationship between renewable energy
investments and financial performance; however, the strength and direction of
this relationship has not been thoroughly assessed by scholars. Secondly, it
has shown that policy plays a paramount role in increasing investors’
confidence and decreasing the investment risk. At the same time, however, it
has higlighted that policy can be perceived as an additional risk factor for
investors. Thirdly, it has emphasized that cognitive factors have an impact on
the decision to invest in renewables. Therefore policy effectiveness is critically
dependent upon its impact on investors’ behaviours. Distorted risk perceptions
may lead to suboptimal decisions and represent additional barriers to
renewable energy market growth. This issue has relevant implications both at
a micro and macroeconomic level. Under a microeconomic perspective,
biased perceptions vis-à-vis renewable energy technologies can lead
investors to overlook promising opportunities in a sector characterized by a
significant growth potential. At a macroeconomic level, if renewable energy
156
policies fail to understand the behavioural context in which investors make
decisions, they will not be able to leverage the needed amount of capital to
foster the transition toward more sustainable energy paths. In order to
maximize the impact of future energy policies, public regulators therefore
need to get a better understanding of how investors behave, and of how they
take their decisions, particularly in regards to the key psychological factors
that may influence their actions.
Yet, there is a surprising lack of rigorous empirical studies examining these
issues in the energy policy and strategic management literature. This work
represents one of the first attempts to fill this gap.
Drawing upon a comprehensive literature review encompassing energy policy,
finance, management and cognitive psychology studies, a conceptual model
has been put forth in order to address the two main research questions of the
doctoral work. The model examines whether behavioural factors have a
measurable influence on the decision to invest in renewable energy projects
and whether, in turn, the share of renewable energy in the portfolio that results
from these decisions is reflected into the portfolio performance. Three main
categories of behavioural factors have been included as independent
variables in the model: i) a priori-beliefs, ii) policy preferences and, iii)
attitudes toward technological risk. As for the dependent variables, these are
the renewable energy share in the portfolio and the investment performance.
The methodological approach followed is rigorous since it combines cognitive
psychology approaches with quantitative measurement techniques, therefore
allowing to gather statistically relevant results. Two main measurement tools
were adopted, serving specific research purposes. More specifically, ACA was
used to investigate investors’ preferences over renewable energy policy
characteristics, whereas the multivariate regression analysis had the goal to
measure the statistical relationships between the dependent and independent
variables. The conceptual model developed has proven to be able to capture
the elements influencing the investment decision making process in the
renewable energy field, as well as the strength and the direction of the
relationship between variables, therefore providing clear answers to the
research problem.
In particular, it has demonstrated that cognitive factors have a measurable
influence on the decision to invest in renewables. The multivariate regression
results have revealed that a priori beliefs on the technical effectiveness of the
investment opportunities play a much more important role than market
efficiency beliefs in driving investments. This implicitly suggests that agents
consider the proven reliability of a technology as a necessary condition for
157
investing in it, whilst they believe that market inefficiencies can be corrected
through the adoption of appropriate policy instruments.
The analysis of the performance model has found that higher shares of
renewable energy technologies in the investment portfolio are associated with
a slightly higher performance relative to direct competitors, thus answering the
second research question.
The results have also depicted a group of investors with relatively short
investment horizons, who have a strong preference for policies that provide
high levels of financial incentives for a limited amount of time over policies that
guarantee a moderate but stable support for a longer time. Furthermore, a
tendency to invest in radically new technologies does not translate
automatically into a higher share of renewable energies in the portfolio.
The research has also revealed that the majority of survey respondents have
rather limited experience in the renewable energy sector. Although this is not
surprising, since the renewable energy market has evolved only quite recently,
the lack of accumulated experience implies that financiers might fail to
analyse investment opportunities in a proper way. In fact, common financial
rules of thumb are not necessarily suited to guide investment decisions in a
highly specialised field as the renewable energy market. This might lead to
overlook promising opportunities by privileging short term returns instead of
embarking in projects that tend to guarantee superior returns over a longer
term horizon. This issue has profound implications also for policy making,
since time horizons of relevance to investors are not necessarily those which
are socially optimal. This is a very important point to be taken into account
when investigating the relationship between policy and investment.
Furthermore, the ACA results reveal that investors are primarily driven by total
remuneration opportunities. Therefore reasonably high levels of the premium
incentive should be given to renewable energy projects in order to ensure that
investments will be undertaken. Feed-in tariffs are the most favoured policy
instruments. Investors also ask for continuity of support, over a timeframe
longer than 20 years. Other aspects like the administrative process and social
acceptance seem to represent less of an issue. However, the segmentation
analysis has emphasized some important differences across the sample.
Results indicate that policies are perceived differently, depending on the
investor category, the geographical location and the type of renewable energy
technology.
158
The present study makes a contribution to management, energy policy and
behavioural finance literature, and has some important implications for
managerial practice.
Firstly, by providing statistically robust results on the positive relationship
between renewable energy investments and financial performance, the
research corroborates the existing literature by providing empirical evidence
on an under-investigated topic so far.
Secondly, the research had the purpose to assess investors’ attitudes toward
renewable energy policies. Since renewable energy markets are regulated
under several policy schemes, understanding which policy characteristics
influence the most the likelihood to invest in a renewable energy project by
reducing the perceived risk associated to the investment can help policy
makers design more effective policy instruments to support the market
deployment of renewables.
Furthermore, the development of a conceptual model which incorporates
cognitive and behavioural elements into the analysis of the renewable energy
investment decision making process is an important methodological
contribution which produces a more accurate description of the relationship
between policies and investment.
The analysis of the influence of behavioural factors on the investment decision
making process represents an important contribution also to behavioural
finance. By providing empirical findings on how and to what extent investment
behaviours in the renewable energy sectors deviate from the expectations of
traditional finance theories, the present works help corroborate this stream of
research.
The study also contributes to the theory of social acceptance of renewable
energy innovation. As observed by Wüstenhagen et al. (2007), while factors
influencing socio-political and community acceptance are increasingly
recognized as being important in the understanding of policy effectiveness,
market acceptance has received less attention so far. By investigating
investors’ acceptance of renewable energy policies, the present research
contributes to fill this gap.
Finally, the results appear also relevant for practitioners in the renewable
energy market. A priory beliefs and cognitive biases create additional risk
elements that restrain the likelihood of raising capital for clean energy
investments. The analysis of these elements as opposed to more rational risk
factors can help investors get a more balanced view of policy risks and
opportunities in this promising business sector.
159
Like most research, also this study is not exempt from limitations.
As far as the ACA analysis is concerned, the current experiment included five
policy attributes for a total number of 16 attribute levels. Additional research is
recommended by using the same number of levels across attributes in order
to accurately assess to what extent a higher number of levels influence the
attribute importance.
As far as the regression model is concerned, a methodological shortcoming is
related to the fact that the variables used in the portfolio performance model
are self assessed and measured by means of a three-point likert scale. This
choice was dictated by the results of the pre-test, which revealed that
investors were reluctant to disclose any performance-related information
beyond a mere first order assessment of their ranking with respect to peers
(below, in line with, above). Although the analysis excluded the presence of
common method variance problems and the performance equation was reestimated through a multinomial logit model, the use of objective, quantitative
measures of performance would have been ideal and remains necessary to
further validate the findings.
Additionally, the presence of a reverse causal relationship between
investment performance and the share of renewables in the investment
portfolio cannot be excluded a priori. Over time, it can be expected that
rational investors who obtained above-average returns by adding renewables
to their portfolios, will also tend to increase the share of renewables in their
portfolios in the next investment round. Therefore it would be interesting to
test the relationship between investment performance at time t and the
renewable energy share in the investment portfolio at time t+1.
Fully disentangling this reverse causality would require the availability of
longitudinal data, which, unfortunately, were not available for this study.
Addressing some of these issues in follow-up research would bring both
methodological and theoretical added values. In this respect, additional
analysis through face to face interviews with a selected sample of
stakeholders might help overcome some of the barriers encountered during
the questionnaire survey, thus perhaps facilitating the obtaining of additional
information on performance data. It is also worth adding that many companies
investing in the renewable energy domain have a relatively short history. This
limits the possibility to get performance data on a relatively long period of time.
Further research could therefore restrict the focus on a smaller sub-sample
with a longer experience in the renewable energy investment business. On
the one hand, a smaller sub-sample would allow performing more in-depth
case studies; on the other hand, the need to obtain objective performance
160
measurements would imply also the possibility to get access to subscriptionbased magazines and directories in order to cross-check the information
provided by companies against external sources.
The present study is based on a specific empirical context. Results may be
difficult to generalize because the sample is skewed toward venture capitalists
and private equity funds. Additional research is recommended with a wider
sample of investor categories. Ideally, it would be worth conducting investorspecific studies that can be compared to each other in order to gather more
accurate insights on the investment decision making process and on policy
perceptions of different category of stakeholders.
Furthermore, the relatively limited sample size did not allow for a better
differentiation among the various types of renewable energy investments. The
survey included investments in a wide range of renewable energy
technologies characterized by different degrees of innovativeness and risk.
Clearly, some of the phenomena observed may be technology-dependent and
require further investigation. Therefore, technology-specific studies would be
also very appropriate. In this respect, four renewable energy technologies
appear of particular interest for specific reasons: biofuels, solar PV, solar CSP
and offshore wind.
As for biofuels, evidence reveals that the market hype has been followed by a
rapid disappointment and the current rate of investments remains below
expectations. Investigating the role of herd behaviours in explaining such
phenomenon would bring very useful indications to policy makers.
Furthermore, the segmentation analysis has found that social acceptance
issues have a greater weight in the decision to invest in biofuels compared to
other technologies.
Solar PV is the fastest-growing renewable energy technology and is expected
to play an important role in future energy scenarios thanks to its flexibility,
technology improvements and the rapid cost reduction. Therefore, a more indepth analysis of how investors perceive solar PV and what they expect from
solar PV policies would be greatly beneficial.
Solar CSP is also expected to give an important contribution to future energy
supply, particularly in the sunbelt countries. For instance, a series of policy
and industrial initiatives in the South Mediterranean region have the objective
to install 20 GW of additional renewable energy capacity by 2020, to be
satisfied mostly by solar CSP. Understanding how investors undertake their
investment decisions in the solar CSP market and what are the main policy
risks associated to this technology is an additional topic for further research.
161
Wind offshore is expected to give a significant contribution to the achievement
of renewable energy targets in the EU by 2020. However, wind offshore
presents peculiar challenges with respect to interconnection and grid
integration. This makes this technology an additional relevant case study.
Finally, the study was restricted to the European region. Results might be
different in other empirical contexts, due to the different policy frameworks in
place and other country specific boundary conditions, as well as investors’
attitudes. In order to capture these differences and get a better understanding
of investors’ behaviours, it would be interesting to conduct similar surveys in
other world regions. Particularly interesting case studies would be countries
like the United States, China, India and Brazil which are currently triggering
investments in various segments of the renewable energy market. Such
additional studies would provide relevant insights for national and international
investors, as well as for policy makers and international institutions.
Summarising, the main recommendations deriving from the study results are
that more targeted policies should be designed, which must take into account
the degree of renewable energy technology maturity, country-specific
conditions and that serve the specific needs and concerns expressed by
different categories of operators. Using the term coined by Hamilton (2009),
more “investment-grade” policies are needed in order to stimulate scaled-up
financial resources in the renewable energy field. At the same time, policies
can also contribute to a change in the way financial actors approach the
renewable energy business, by fostering a longer term perspective. As
pointed out by Grubb “technological advances, and in some cases
breakthroughs, are certainly needed: but the revolution required is one of
attitudes” (Grubb, 1990, page 716). Therefore, if the share of renewables in
the global energy mix is to be increased, significant innovations are needed
not only in the technical field, but also in the social and institutional context
(Krewitt et al., 2007).
While working on the current dissertation, a series of suggestions for further
research have come out. In particular, it became evident that so far
government spending in the renewable energy sector has been directed
mainly to renewable energy deployment. As a consequence, most literature
studies have concentrated on evaluating the effectiveness of various policy
mechanisms in leading to a widespread diffusion of renewable energy
technologies. Much less attention has been devoted to the assessment of
policy effectiveness in bringing renewable energy technology innovation and
development, as also highlighted by Nemet and Kammen (2007). This seems
to be a research priority. In fact, in order for renewables to become
162
mainstream, both policy makers and financiers need to better understand the
main issues and identify the appropriate solutions to overcome the valley of
death. This would accelerate cost reduction, bridge the current
competitiveness gap and lead to the creation of a self-sustaining market for
renewables.
Within this framework, studies incorporating the behavioural finance
perspective in the analysis of technology-push policies would bring significant
added value.
Another aspect which became clearer while working on the present project is
that the success of a sustainable energy technology does not depend only on
the effectiveness of dedicated policies, but on a more complex mix of
ingredients, which include also the existence of a dynamic industrial
environment, the implementation of additional measures to support innovation,
and public/private partnerships in the R&D area.
In a previous paper (Menichetti, 2005) I had already highlighted the positive
role of spillover effects in nurturing the growth of PV. The spillover approach
seems to be particularly appropriate to study renewable energy technology
innovation dynamics. Conducting further research by analyzing the underlying
factors which can explain why some sustainable technologies have become
more successful than others (e.g: PV vs. hydrogen) is a clear priority in my
next works.
163
10 References
Akerlof, G.A. and Yellen, J.L. (1987). Rational Models of Irrational Behavior.
The American Economic Review, Vol. 77(2), 137-142.
Allcott, H. and Mullainathan, S. (2010). Behavior and Energy Policy. Science,
Vol. 327(5970), 1204-1205.
Allen, D.R. (2004). Customer Satisfaction Research Management: A
comprehensive guide to integrating customer loyalty and satisfaction metrics
in the management of complex organizations. American Society for Quality,
Quality Press, Milwaukee.
Alriksson, S. and Öberg, T. (2008). Conjoint analysis for environmental
evaluation – A review of methods and applications. Environmental Science
and Pollution Research 15(3), 244-257.
Álvarez-Farizo, B. and Hanley, N. (2002). Using Conjoint Analysis to Quantify
Public. Preferences Over Environmental Impacts of Wind Farms. An Example
from Spain. Energy Policy, Vol. 30(2), 107-116.
Andersson, L. M. and Bateman, T.S. (1997). Cynicism in the Workplace:
Some Causes and Effects. Journal of Organizational Behavior, Vol. 18(5),
449-469.
Awerbuch, S. (2003a). Determining the real cost: why renewable power is
more cost-competitive than previously believed. Renewable Energy World,
Vol. 6(2).
Awerbuch, S. (2003b). Do renewables cost more? Shifting the grounds of
debate. Presentation at the Sonderborg conference on Renewable Energy –
Renewable Energy in the Market: New Opportunities, Sondenborg, Denmark,
17-19 September 2003.
Awerbuch, S. (2000). Investing in photovoltaics: risk, accounting and the value
of new technology. Energy Policy, Vol.28, 1023-1035.
Awerbuch, S. (1996). Capital budgeting, technological innovation and the
emerging competitive environment of the electric power industry. Energy
Policy, Vol. 24(2), 195-202.
Awerbuch, S., and Berger, M. (2003). Applying portfolio theory to EU
electricity planning and policy-making, IEA working paper EET/2003/03, Paris,
February 2003.
Awerbuch, S., and Sauter, R. (2006). Exploiting the oil-GDP effect to support
renewable deployment. Energy Policy, Vol.34, 2805-2819.
Backhaus, K., Erichson, B., Plinke, W., Weiber, R. (1996). Multivariate
Analysemethoden - Eine anwendungsorientierte Einführung, 7 ed. Berlin:
Springer-Lehrbuch.
164
Banfi, S., Farsi, M., Filippini, M., Jakob, M. (2008). Willingness to pay for
energy-saving measures in residential buildings. Energy Economics, Vol.
30(2), 503-516.
Barberis, N. and Thaler, R. (2003). A Survey of Behavioral Finance. In:
Handbook of the Economics of Finance, edited by G.M. Constantinides, M.
Harris and R. Stulz, Elsevier.
Barberis, N., Huang, M. and Santos, T. (1999). Prospect Theory and Asset
Prices, Mimeo, University of Chicago.
Barberis, N., Shleifer, A. and Vishny, R. (1998). A model of investor
sentiment. Journal of Financial Economics, Vol. 49, 307-345.
Batsell, R. R. and Lodish, L. M. (1981). A model and measurement
methodology for predicting individual consumer choice. Journal of Marketing
Research, Vol. 18, 1-12.
Bazilian, M. and Roques, F. (Eds.), 2008. Analytical Methods for Energy
Diversity and Security. Portfolio Optimization in the Energy Sector: A Tribute
to the Work of Dr. Shimon Awerbuch. Elsevier, Amsterdam.
Bendoly, E. (2006). Incorporating behavioural theory in OM empirical models.
Journal of Operations Management, Vol. 24(6), 735-736.
Beretti, A., Grolleau, G. and Mzoughi, N. (2009). How cognitive biases can
affect the performance of eco-labeling schemes. Journal of Agricultural and
Food Organization, Vol. 7(2), Article 10.
Bergek, A. and Jacobsson, S. (2010). Are tradable green certificates a costefficient policy driving technical change or a rent-generating machine?
Lessons from Sweden 2003-2008. Energy Policy, Vol. 38(2010), 1255-1271.
Bernantzi, S. and Thaler, R.H. (1995). Myopic Loss Aversion and the Equity
Premium Puzzle, Quarterly Journal of Economics, 110(1), 73-92.
Birley, S., Muzyka, D. and Hay, M. (1999). Management Buyout: Perception of
Opportunity. A Research Note. Journal of Management Studies, Vol. 36, 109122.
Birol, F. (2010). Introductory remarks and presentation, IEA World Energy
Outlook workshop on “Outlook for renewable energy”, Paris, 26 April 2010.
Black, T.R. (1999). Doing quantitative research in the social sciences: An
integrated approach to research design, measurement and statistics. London,
Sage, 2-58.
Blok, K. (2006). Renewable energy policies in the European Union. Energy
Policy, Vol. 34(3), 251-255.
Bloomberg NEF - New Energy Finance (2010). Energy: Week in Review, Vol.
6(29), 30 March-12 April 2010.
Bloomberg NEF – New Energy Finance (2009a). Clean Energy Investment in
the Economic Downturn, Rome, June 2009.
165
Bloomberg NEF – New Energy Finance (2009b). Energy: Week in Review,
6(9), 27 October – 2 November 2009.
Bromiley, P. (2005). The Behavioral Foundations of Strategic Management,
Blackwell Publishing, Malden, MA, 2005.
Buckley, M.R., Cote, J.A., Comstock, S.M. (1990). Measurement errors in
behavioral sciences: The case of personality/attitude research. Educational
and Psycological Measurement, Vol. 50(3), 447–474.
Bürer, M.J. (2008). Public Policy and Clean Energy Private Equity Investment,
doctoral dissertation nr. 3421, University of St. Gallen, Graduate School of
Business, Economics, Law and Social Sciences.
Bürer, M.J. and Wüstenhagen, R. (2009). Which renewable energy policy is a
venture capitalist’s best friend? Empirical evidence from a survey of
international cleantech investors. Energy Policy, Vol. 37(12), 4997-5006.
Bürer, M.J. and Wüstenhagen, R. (2008a). Cleantech venture investors and
energy policy risk: an exploratory analysis of regulatory risk management
strategies. In: R.Wustenhagen, J. Hamschmidt, S. Sharma and M. Starik
(Eds), Sustainable Innovation and Entrepreneurship, Edward Elgar Publishing
(2008), 290-309.
Bürer, M.J. and Wüstenhagen, R. (2008b). Effectiveness of renewable energy
and climate policies: an empirical analysis of cleantech investor preferences,
draft working paper.
Butler, L., Neuhoff, K. (2004). Comparison of feed in tariff, quota and auction
mechanisms to support wind power development, CMI/DAE Working paper
70.
Carroll, J.D. (1969). Categorical conjoint measurement.
Mathematical Psychology, Ann Arbor, MI, August.
Meeting
of
Cohen, M., Fenn, S., and Naimon, J. (1995). Environmental and financial
performance: Are they related? Working paper, Vanderbilt University,
Nashville, TN.
Connor, P.M. (2003). UK renewable energy policy: a review. Renewable and
Sustainable Energy Reviews, Vol. 7(2003), 65-82.
Contaldi, M., Gracceva, F., Tosato, G. (2007). Evaluation of green-certificates
policies using the MARKAL-MACRO-Italy model. Energy Policy, Vol. 35(2),
797-808.
Couture, T. and Gagnon, Y. (2010). An analysis of feed-in tariff remuneration
models: Implications for renewable energy investment. Energy Policy, Vol.
38(2), 955-965.
Christiansen, A.C. (2001). Technological change and the role of public policy:
an analytical framework for dynamic efficiency assessments. Report of the
Fridtjof Nansen Institute, 2001.
166
Culhane, T.H. (2008). Solar energy in Egypt: A question of behavioral
economics, in: Tarek H. Selim (Ed.): Egypt Energy and the Environment –
Critical Sustainability Perspectives, chapter 8, 93- 110, The American
University in Cairo, June 2008.
Cutler, D., Poterba, J. and Summers, L. (1991). Speculative dynamics.
Review of Economic Studies, Vol. 58, 529-546.
Daniel, K., Hirshleifer, D. and Subrahmanyam (1998). Investor psychology
and security market under- and overreactions. Journal of Finance, Vol. 53,
1839-1885.
DeBondt, W. and Thaler, R. (1985). Does the stock market overreact? Journal
of Finance, 40, 793-808.
del Río González, P. (2008). Ten years of renewable electricity policies in
Spain: An analysis of successive feed-in tariff reforms. Energy Policy, Vol.
36(8), 2917-2929.
Deutsche Bank (2009). Infrastructure Investments in Renewable Energy,
September 2009 (retrieved from: http://www.dbcca.com/research).
Deutsche Bank (2008). Investing in Climate Change 2009 – Necessity and
Opportunity in Turbulent Times, October 2008 (retrieved from:
http://dbadvisors.com/climatechnage).
Dinica, V. (2006). Support systems for the diffusion of renewable energy
technologies – an investor perspective. Energy Policy, Vol. 34(2006), 461480.
Dowell G, Hart S, Yeung B. (2000). Do corporate global environmental
standards create or destroy market value? Management Science, Vol. 46(8),
1059-1074.
Dufhues, T., Heidhues, F. and Buchenrieder, G. (2004). Participatory Product
Design by Using Conjoint Analysis in the Rural Financial Market of Northern
Vietnam. Asian Economic Journal, Vol. 18(1), 81-114.
EC- European Commission (2008). The support of electricity from renewable
energy sources, accompanying document to the proposal for a directive of the
European Parliament and the Council on the promotion of the use of energy
from renewable sources, Commission Staff Working Document SEC,
Brussels, 23 January 2008.
EC- European Commission (2005). Communication from the Commission –
The support of electricity from renewable energy sources, COM(2005) 627
final, Brussels, 7 December 2005.
EC- European Commission (1999). Electricity from renewable energy sources
and the internal electricity market, Commission working document SEC (1999)
470 final, 13 April 1999.
ECF (2010). Roadmap 2050 – Practical guide to a prosperous, low-carbon
Europe, Vol. 1 and 2, April 2010.
167
Ecofys (2008) – de Jager, D. and Rathmann, M. Policy Instrument Design to
Reduce Financing Costs in Renewable Energy Technology Projects. Final
draft for the IEA Implementing Agreement on Renewable Energy Technology
Deployment, Utrecht.
Einhorn, H.J. and Hogarth, R.M. (1986). Decision Making under Ambiguity.
Journal of Business, Vol. 59(4), S225-S250.
EPIA (2010). Global Market Outlook for Photovoltaics until 2014, European
Photovoltaic Industry Association, Brussels, 2010.
EPIA (2009). Global Market Outlook for Photovoltaics until 2013, European
Photovoltaic Industry Association, Brussels, 2009.
EREC (2010). Re-thinking 2050 - A 100% Renewable Energy Vision for the
European Union, European Renewable Energy Council (EREC), Brussels,
2010.
Espey, S. (2001). Renewables portfolio standard: a means for trade with
electricity from renewable energy sources?. Energy Policy, Vol. 29(2001),
557-566.
EWEA (2010). Wind in Power – 2009 European statistics, February 2010.
Fama, E. (1998). Market efficiency, long-term returns, and behavioral finance.
Journal of Financial Economics, Vol. 49, 283-306.
Fama, E. (1970). Efficient Capital Markets: A Review of Theory and Empirical
Work. Journal of Finance, Vol. 25, 383-417.
Fama, E. (1965). The Behavior of Stock Market Prices. Journal of Business,
Vol. 38, 34-105.
Farsi, M. (2010) Risk aversion and willingness to pay for energy efficient
systems in rental apartments. Energy Policy, Vol.38(6), 3078-3088.
Fouquet, D. and Johansson, T.B. (2008). European renewable energy policy
at crossroads – Focus on electricity support mechanisms. Energy Policy, Vol.
36(2008), 4079-4092.
Franke, N., Gruber, M., Harhoff, D., Henkel, J. (2006). Venture Capitalists’
evaluations of start-up teams: A discrete choice analysis, working paper,
January 2006 (retrieved from: http://www.tim.wi.tum.de/paper/Franke_et_alEvaluations_of_start-up_teams_2006.pdf)
Frankl, P. and Philibert, C. (2009). Critical role of renewable energy to climate
change mitigation, presentation held at the COP15 IEA Day Side Event,
Copenhagen, 16 December 2009.
Frankl, P., Menichetti, E., Raugei, M. (2006): Final Report on Technical Data,
Costs and Life Cycle Inventories of PV Applications. NEEDS Project
Deliverable No. 11.2 – RS Ia. Retrieved from: http://www.needsproject.org/RS1a/RS1a%20D11.2%20Final%20report%20on%20PV%20techn
ology.pdf.
168
Friedman, M. (1953). The case for Flexible Exchange Rates. In: Essays in
Positive Economics, University of Chicago Press, Chicago.
Fritz-Morgenthal, S., Greenwood, C., Menzel, C., Mironjuk, M., SonntagO’Brien, V. (2009). The global financial crisis and its impact on renewable
energy finance, UNEP SEFI publication, April 2009.
Froot, K. and Dabora, E. (1999). How are stock prices affected by the location
of trade?. Journal of Financial Economics, Vol. 53, 189-216.
Gino, F. and Pisano, G. (2008). Toward a Theory of Behavioral Operations.
Manufacturing & Service Operations Management, Vol. 10(4), 676-691.
Glaser, B.G. and Strauss, A.L. (1967). The discovery of grounded theory:
Strategies for qualitative research, Chicago: Aldine Publishing Company.
Gompers, P. and Lerner, J. (2001). The venture capital revolution. Journal of
Economic Perspectives, Vol. 15(2), 145-168.
Green, P. and Rao, V.R. (1971). Conjoint Measurement for Quantifying
Judgmental Data. Journal of Marketing Research, Vol. 8, 355-363.
Green, P.E. and Rao, V.R. (1969). Nonmetric approaches to multivariate
analysis in Marketing. Working paper, Wharton School, University of
Pennsylvania.
Green, P.E. and Srinivasan, V. (1990). Conjoint Analysis in Marketing: New
Developments With Implications for Research and Practice. Journal of
Marketing, Vol. 54(4), 3-19.
Green, P. and Srinivasan, V. (1978). Conjoint analysis in consumer research:
Issues and outlook. Journal of Consumer Research, Vol. 5(2), 103-123.
Green, P. E. and Wind, Y (1975). New Way to Measure Consumer
Judgments. Harvard Business Review, Vol.53 (July-August), 107-117.
Green, P. E. and Wind, Y. (1973). Multiattribute Decisions in Marketing: A
Measurement Approach. Hinsdale, IL: The Dryden Press.
Greenpeace International and EREC (2008). Energy [r]evolution – A
Sustainable Global Energy Outlook, Greenpeace International and European
Renewable Energy Council (EREC), Amsterdam/Brussels, October 2008.
Grubb, M.J. (2004). Technology Innovation and Climate Policy, an overview of
issues and options. Keio Economic Studies, Vol. 41(2), 103-132.
Grubb, M.J. (1990). The Cinderella options. A study of modernized renewable
energy technologies. Part 1-A technical assessment. Energy Policy, Vol.
18(6), July-August 1990, 525-542.
Gustafsson, A., Herrmann, A., Huber, F. (2007), Conjoint analysis as an
instrument of market research practice. In: Gustafsson, A., Herrmann, A.,
Huber, F. (Eds), Conjoint Measurement: Methods and Applications, Fourth
Edition, Springer, Berlin, 5-47.
169
GWEC (2010). Global Wind Report 2009, Global Wind Energy Council,
Brussels, March 2010.
GWEC (2009). Global Wind Report 2008, Global Wind Energy Council,
Brussels, March 2009.
Haas, R., Eichhammer, W., Huber, C., Langniss, O., Lorenzoni, A., Madlener,
R., Menanteau, P., Morthorst, P.E., Martins, A., Oniszk, A., Schleich, J.,
Smith, A., Vass, Z., Verbruggen, A. (2004). How to promote renewable energy
systems successfully and effectively. Energy Policy, Vol. 32(2004), 833-839.
Hair, J.F.Jr., Anderson, R.E., Tatham, R.L. Black, W.C. (1998). Multivariate
Data Analysis, Fifth Edition, Upper Saddle River, New Jersey: Prentice Hall.
Hamilton, K. (2009). Unlocking Finance for Clean Energy: The need for
“investment grade” policy. Chatham House, Energy, Environment and
Development programme paper 09/04, London, December 2009.
Hardyck, C.D. and Petrinovich, L.F. (1976). Introduction to Statistics for the
Behavioral Sciences, Second Edition, Philadelphia: Saunders.
Hart, S. and Ahuja, G. (1996). Does it pay to be green? An empirical
examination of the relathionship between emission reduction and firm
performance. Business Strategy and the Environment, Vol. 5, 30-37.
Hasan, R. (2006). A Research Report on Ethanol Investment: The Ethanol
th
Hype, Quantified. Altenews, July 17 2006.
Hausman, J.A. (1978). Specification Tests in Econometrics. Econometrica,
Vol. 46(6), 1251-1271.
Heinzle, S. and Wüstenhagen, R. (2009). Consumer survey on the new format
of the European Energy Label for televisions - Comparison of a "A-G closed"
versus a "beyond A" scale format, working paper n.3 within the project
“Soziale, ökologische und ökonomische Dimensionen eines nachhaltigen
Energiekonsums in Wohngebäuden”, St. Gallen, August 2009.
Held, A., Haas, R., Ragwitz, M. (2006). On the Success of Policy Strategies
for the Promotion of Electricity from Renewable Energy Sources in the EU,
report produced for DG Tren.
Hitt, M.A. and Tyler, B.B. (1991). Strategic Decision Models: Integrating
Different Perspectives. Strategic Management Journal, Vol. 12(5), 327-351.
Hopkins, D.S.P., Larréché, J.C., Massy, W.F. (1977). Constrained
optimization of a University administrator’s preference function. Management
Science, Vol. 24, 365-377.
Huber, F., Herrmann, A., Gustafsson, A. (2007). On the influence of the
Evaluation Methods in Conjoint Design – Some Empirical Results. In:
Gustafsson, A., Herrmann, A., Huber, F. (Eds), Conjoint Measurement:
Methods and Applications, Fourth Edition, Springer, Berlin, 93-112.
170
Huber, G.P. and Power, D.J. (1985). Retrospective reports of strategic-level
managers: Guidelines for increasing their accuracy. Strategic Management
Journal, Vol. 6, 171-180.
Hvelplund, F. (2005). Renewable energy: Political prices or political quantities,
in: Volkmar Lauber (Ed.): Switching to renewable power. A framework for the
st
21 century, London: Earthscan, 228-245.
IEA (2010a). Energy Technology Roadmap – Solar Photovoltaic Energy,
OECD/IEA, Paris, 2010
IEA (2010b). Energy Technology Roadmap – Concentrating Solar Power,
OECD/IEA, Paris, 2010
IEA (2009a). World Energy Outlook, OECD/IEA, Paris, 2009.
IEA (2009b). Energy Technology Roadmaps – Charting a low-carbon energy
revolution, OECD/IEA, Paris, 2009.
IEA (2009c). Renewable Energy Essentials – Concentrating Solar Power,
OECD/IEA, Paris, 2009.
IEA (2009d). From 1st- to 2nd-Generation Biofuel Technologies - An overview
of current industry and RD&D activities, OECD/IEA, Paris, 2009.
IEA (2009e). IEA Scoreboard 2009 – 35 Key Energy Trends over 35 Years,
IEA (2009f). The Impact of the Financial and Economic Crisis on Global
Energy Investment, OECD/IEA, Paris, 2009.
IEA (2008a). Energy Technology Perspectives 2008, OECD/IEA, Paris 2008.
IEA (2008b). Deploying Renewables – Principles for Effective Policies,
IEA (2007). Climate Policy Uncertainty and Investment Risk, OECD/IEA,
Paris, 2007.
Jacobsson, S. and Johnson, A. (2000). The diffusion of renewable energy
technology: an analytical framework and key issues for research. Energy
Policy, Vol. 28, 625-640.
Jacobsson, S., Bergek, A., Finon, D., Lauber, V., Mitchell, C., Toke, D. and
Verbruggen, A. (2009). EU renewable energy support policy: Faith or facts?.
Energy Policy, Vol. 37(2009), 2143-2146.
Jaeger,S.R., Hedderley, D., MacFie, H.J.H. (2000). Methodological issues in
conjoint analysis: a case study. European Journal of Marketing, Vol.
35(11/12), 1217-1237.
Janssen, W., Ashby, J., Carlier, M. and Castaño, J. (1991). Targeting New
Technology at Consumer Food Preferences in Developing Countries. Food
and Quality Preference, Vol. 3(2), 175–182.
171
Jegadeesh, N. and Titman, S. (1993). Returns to buying winners and selling
losers: implications for stock market efficiency. Journal of Finance, Vol. 48,
65-91.
Johnson, R.M. (2001). History of ACA. Sawtooth Software Research paper
series, Sequim, WA.
Johnson, R.M. (1987). Adaptive Conjoint Analysis. Sawtooth Software
Conference Proceedings, Ketchum: Sawtooth Software, July, 253-265.
Kahneman, D. (2003). Maps of Bounded Rationality: Psychology for
Behavioral Economics. The American Economic Review, Vol. 93(5), 14491475.
Kahneman, D. and Tversky, A. (1979). Prospect Theory: An Analysis of
Decision Under Risk. Econometrica, Vol. 47(2), 263-292.
Kaplan, R. (1986). Must CIM be justified on faith alone? Harvard Business
Review, Vol. 64(2), 87-95.
Kerlinger, F. (1986). Foundations of Behavioral Research, Third Edition,
Orlando, FL: Harcourt Brace & Company.
King, A.A. and Lenox, M.J. (2001). Does it really pay to be green? An
empirical study of firm environmental and financial performance. Journal of
Industrial Ecology, Vol. 5(1), 105-116.
King, W. C., Hill, A. and Orme, B. (2004). The ‘Importance’ Question in ACA:
Can It Be Omitted?. Sawtooth Software Conference Proceedings, Sequim,
WA: Sawtooth Software.
Krantz, D.H. (1964). Conjoint measurement: The Luce-Tukey axiomatization
and some extensions. Journal of Mathematical Psychology, Vol. 1, 248-277.
Krewitt, W., Simon, S., Graus, W., Teske, S., Zervos, A. and Schäfer, O.
(2007). The 2° C scenario – A sustainable world energy perspective. Energy
Policy, Vol. 35, 4969-4980.
Kruskal, J.B. (1965). Analysis of factorial experiments by estimating monotone
transformations of the data. Journal of the Royal Statistical Society, Series B,
Vol.27, 251-263.
Landström, H. (1998). Informal investors as entrepreneurs. Technovation, Vol.
18(5), 321-333.
Langniss, O. (Ed.) (1999). Financing Renewable Energy Systems, DLR,
Stuttgart.
Lauber, V. (2004). REFIT and RPS: options for a harmonised Community
framework. Energy Policy, Vol. 32(2004), 1405-1414.
Leggett, J. (2009). Fossil future, solar future: Implications for Sociey. Power
Point presentation, Tage der Technik 2009, St. Gallen.
172
Leleux, B. and Surlemont, B. (2003). Public versus private venture capital:
seeding or crowding out ? A pan-European analysis. Journal of Business
Venturing, Vol. 18, 81-104.
Lenk, P.J., DeSarbo, W.S., Green, P.E., Young, M.R. (1996). Hierarchical
Bayes Conjoint Analysis: Recovery of Partworth Heterogeneity from Reduced
Experimental Designs. Marketing Science, Vol. 15(2), 173-191.
Lesser, J.A. and Su, X. (2008). Design of an economically efficient feed-in
tariff structure for renewable energy development. Energy Policy, Vol.
36(2008), 981-990.
Liang, X. and Reiner, D. (2009). Behavioral issues in financing low carbon
power plants. Energy Procedia, Vol. 1(1), 4495-4502.
Liebreich, M. (2010). No title. Power Point presentation, IEA Renewable
Energy Working Party Workshop: “Renewables, from Cinderella options to
mainstream energy solution”, Paris, 15 March 2010.
Liebreich, M. (2009). Feed-in tariffs: Solutions or time-bomb? New Energy
Finance Monthly Briefing, 5(28), August 2009, 1-3.
Lind, R.C. (1982). A primer on the major issues relating to the discount rate
for evaluating national energy options. In: Lind, R.C., Arrow, K. et al. (eds.),
Discounting for Time and Risk in Energy Policy. Johns Hopkins University
Press, Baltimore (published for Resources for the Future, Washington, DC).
Lintner, J. (1965). The Valuation of Risky Assets and the Selection of Risky
Investments in Stock Portfolios and Capital Budgets. Review of Economics
and Statistics, Vol. 47 (1), 13-37.
Lipp, J. (2007). Lessons for effective renewable electricity policy from
Denmark, Germany and the United Kingdom. Energy Policy, Vol. 35(11),
5481-5495.
Loch, C.H. and Yaozhong, W. (2005). Behavioral Operations Management.
Foundations and Trends in Technology, Information & Operations
Management, Vol. 1(3), 121-232.
Lorenzoni, A. (2003). The Italian Green Certificates market between
uncertainty and opportunities. Energy Policy, Vol. 31 (2003), 33-42.
Louviere, J.J. and Woodworth, G.G. (1983). Design and Analysis of Simulated
Consumer Choice or Allocation Experiments: An Approach Based on
Aggregated Data. Journal of Marketing Research, Vol. 20, 350-367.
Lovallo, D. and Sibony, O. (2010). The case for behavioral strategy. McKinsey
Quarterly, March 2010.
Luce, D. and Tukey, J. (1964). Simultaneous Conjoint Measurement: A New
Type of Fundamental Measurement. Journal of Mathematical Psychology,
Vol. 1, 1-27.
173
Lüthi, S. and Wüstenhagen, R. (2009). The Price of Policy Risk – Empirical
Insights from choice Experiments with European Photovoltaic Project
Developers. Paper prepared for the final conference of the EU INCO project
DISTRES, Nicosia, 11-12 December 2009.
Madlener, R. and Stagl, S. (2005). Sustainability-guided promotion of
renewable electricity generation. Ecological Economics, Vol. 53(2), 147-167.
Malkiel, B.G. (1995). Returns from Investing in Equity Mutual Funds. Journal
of Finance, Vol. 50(2), 549-572.
March, J. (1994). A Primer on decision making: How decisions happen. The
Free Press, New York.
March, J.G. and Shapira, Z. (1987). Managerial Perspectives on Risk and
Risk Taking, Management Science, Vol. 33(1), 1404-1418.
Marshall, P. and Bradlow, E.T. (2002). A Unified Approach to Conjoint
Analysis Models. Journal of the American Statistical Association, Vol. 97(459),
674-682.
Markowitz, H. (1952). Portfolio selection. The Journal of Finance, Vol. 7, 7791.
Maxim, P.S. (1999). Quantitative research methods in the social sciences,
New York: Oxford University Press.
McClain, J.O. and Rao, V.R. (1974). Trade-offs and conflicts in evaluation of
health system alternatives: methodology for analysis. Health Services
Research, Vol. 9(Spring), 35-52.
Mehra, R. and Prescott, E. (1985). The equity premium: a puzzle. Journal of
Monetary Economics, Vol. 15, 145-161.
Menanteau, P., Finon, D., Lamy, M.L. (2003). Prices versus quantities:
choosing policies for promoting the development of renewable energy. Energy
Policy, Vol. 31, 799-812.
Menichetti, E. (2008). Renewable Energy Policy Risk and Investor Behaviour.
Preliminary study for the doctoral dissertation under the Multicultural
Management programme, St. Gallen, 20 May 2008.
Menichetti, E. (2005). The transfer of knowledge and its implications for
technological spillovers, term paper, doctoral seminar on Knowledge, Strategy
and Theories of the Firm, St. Gallen.
Meyer, N.I. (2003). European schemes for promoting renewables in liberalised
markets. Energy Policy, Vol. 31(2003), 665-676.
Meyer, A. et al. (2009). A Copenhagen Climate Treaty, Version 1.0, Karlsruhe,
June 2009.
Miller, E.M. (1977). Risk, Uncertainty and Divergence of Opinion. The Journal
of Finance, Vol. 32(4), 1151-1168.
174
Mitchell, C. and Connor, P. (2004). Renewable energy policy in the UK 19902003. Energy Policy, Vol. 32, 1935-1947.
Mitchell, C., Baucknecht, D., Connor, P.M. (2006). Effectiveness through risk
reduction: a comparison of the renewable obligation in England and Wales
and the feed-in system in Germany. Energy Policy, Vol. 34, 297-305.
Moore, W.L. (1980). Levels of Aggregation in Conjoint Analysis: An Empirical
Comparison. Journal of Marketing Research, Vol. 17, (November), 516–523.
Morthorst, P.E. (2001). Interaction of a tradable green certificate market with a
tradable permits market. Energy Policy, Vol. 29(5), 345-353.
Morthorst, P.E. (2000). The development of a green certificate market. Energy
Policy, Vol. 28(15), 1085-1094.
Mossin, J. (1966). Equilibrium in a Capital Asset Market. Econometrica,
Vol.34, 768-783.
Muñoz, J.I., Sanchez de la Nieta, A.A., Contreras, J., Bernal-Agustin, J.L.
(2009). Optimal investment portfolio in renewable energy: The Spanish case.
Energy Policy, Vol. 37, 5273-5284.
Muñoz, M., Oschmann, V. and Tàbara, J.D. (2007). Harmonization of
renewable electricity feed-in laws in the European Union. Energy Policy, Vol.
35(5), 3104-3114.
Murphy, L.M. and Edwards, P.L. (2003). Bridging the Valley of Death:
Transitioning from Public to Private Sector Financing, May 2003, NREL/MP720-34036, Golden, Colorado.
Muzyka, D., Birley, S. and Leleux, B. (1996). Trade-offs in the Investment
Decisions of European Venture Capitalists. Journal of Business Venturing,
Vol. 11, 273-287.
Neji, L. (2007).Cost development – an analysis based on experience curves,
deliverable 3.3 – RS1a of the EU-funded research project NEEDS – New
Energy Externalities Developments for Sustainability (project nr. 502687),
Lund University.
Nemet, G.F. and Kammen, D.M. (2007). U.S. energy research and
development: Declining investment, increasing need, and the feasibility of
expansion. Energy Policy, Vol. 35(1), 746-755.
Nielsen, L. and Jeppesen, T. (2003). Tradable Green Certificates in selected
European countries – overview and assessment. Energy Policy, Vol. 31(1), 314.
Nielsen, L. and Jeppesen, T. (2000). Green electricity certificates – a
supplement to the flexible mechanisms of the Kyoto protocol. In: Carraro, C.
(Ed.): Efficiency and Equity of Climate Change Policy, Klüwer, The
Netherlands, 221-244 (Chapter 10).
OECD (2008). OECD Environmental Outlook to 2030, Vol. 2008, no. 1, Paris.
175
OECD (2007). OECD Science, Technology and Industry Scoreboard 2007 –
Innovation and Performance in the Global Economy, Paris, November 2007.
OPTRES (2006). Assessment and optimization of renewable support
schemes in the European electricity market, EU IEE project funded by DG
TREN, no. EIE/04/073/s07.38567. D8 report: Analysis of barriers for the
development of electricity generation from renewable energy sources in the
EU-25, Utrecht, May 2006.
Orme, B.K. (2009a). SSI Web 6.6 Manual, 2009 Update, Sawtooth Software
Inc., Sequim, WA.
Orme, B.K. (2009b). Which conjoint method should I use? Sawtooth Software
Research Paper Series, Sawtooth Software Inc., Sequim, WA.
Osborn, R.N. and Jackson, D.H. (1988). Leaders, riverboat gamblers, or
purposeful unintended consequences in the management of complex
dangerous technologies. Academy of Management Journal, Vol. 31, 924-947.
Oschlies, M.K. (2007). A Behavioral Finance Perspective on Sustainable
Energy Investment Decisions, Master Thesis, University of St. Gallen,
Graduate School of Business, Economics, Law and Social Sciences, St.
Gallen, 21 November 2007.
Papadopoulos, A.M. and Karteris, M.M. (2009). An assessment of the Greek
incentives scheme for photovolatics. Energy Policy, Vol. 37(2009), 1945-1952.
Pfluger, B. (2009). Support schemes for renewables in EU, presentation held
at the EWEC pre-conference seminar “Wind Energy – The Facts”, Marseille,
15 March 2009.
Piranfar, H. (2009). Herd mentality and oil prices: implications for
sustainability. International Journal of Global Environmental Issues, Vol. 10 (12), 194-209.
Podsakoff, P.M., MacKenzie, S.B., Lee, J.Y. and Podsakoff, N.P. (2003).
Common method biases in behavioral research: a critical review of the
literature and recommended remedies. Journal of Applied Psychology, Vol.
88(5), 879-903.
Poortinga,W., Steg, L., Vlek, C. and Wiersma, G. (2003). Household
preferences for energy-saving measures: A conjoint analysis. Journal of
Economic Psychology, Vol. 24(1), 49-64.
Porter, M. and van der Linde, C. (1995). Green and competitive. Harvard
Business Review, September-October, 121-134.
Priem, R.L. (1992). An application of metric conjoint analysis for the
evaluation of top managers’individual strategic decision making process: A
research note. Strategic Management Journal, Vol. 13, 143 - 151.
Priem, R.L. and Harrison, D.A. (1994). Exploring strategic judgment: Methods
for testing the assumptions of prescriptive contingency theories. Strategic
Management Journal, Vol. 15(4), 311 - 324.
176
PROGRESS (2008). Promotion and Growth of Renewable Energy Sources
and Systems, EU project funded by DG TREN, no.TREN/D1/422005/SO7.56988, 2008 (http://www.res-progress.eu).
PWC (2010). 100% Renewable Electricity – A Roadmap to 2050 for Europe
and North Africa, Pricewaterhouse Coopers, 2010.
Ragwitz, M. (2010). Effective and efficient long-term oriented RE support
policies. Power Point presentation, IEA Renewable Energy Working Party
Workshop: “Renewables, from Cinderella options to mainstream energy
solution”, Paris, 15 March 2010.
Ragwitz, M. et al. (2009). EmployRES. The Impact of Renewable Energy
Policy on Economic Growth and Employment in the European Union. Final
report, contract no: TREN/D1/474/2006, Karlsruhe, 27 April 2009.
Ragwitz, M. and Huber, C. (2005). Feed-In Systems in Germany and Spain: a
comparison, Fraunhofer Institut für Systemtechnik und Innovationsforschung,
Karlsruhe.
Randolph, T. and Ndung’u, L. (2000). Gender and Transaction Cost: A
Conjoint Analysis of Choice Livestock Health Service among Smallholder
th
Dairy Farmers in Kenya. Background paper for a poster presented at the 24
Congress of the IAAE, 18–24 August 2000, Berlin, Germany.
Rao, V.R. (1977). Conjoint measurement in Marketing analysis. In: Sheth, J.N.
(Ed.) Multivariate Methods for Market and Survey Research, Chicago:
American Marketing Association, 257-286.
Reiche, D. and Bechberger, M. (2004). Policy differences in the promotion of
renewable energies in the EU member states. Energy Policy, Vol. 32(2004),
843-849.
Reinhardt, F.L. (1999). Bringing the environment down to earth. Harvard
Business Review, July-August (114), 149-157.
REN21 (2010). Renewables 2010 Global Status Report, Review Draft, April
th
26 2010.
REN21 (2009). Renewables Global Status Report – 2009 Update, Paris:
REN21 Secretariat and Washington DC: Worldwatch Institute.
Riquelme, H. and Rickards, T. (1992). Hybrid Conjoint Analysis: An Estimation
Probe in New Venture Decisions. Journal of Business Venturing, Vol. 7, 505518.
Rokka, J. and Uusitalo, L. (2008). Preferences for green packaging in
consumer product choices – Do consumers care? International Journal of
Consumer Studies, Vol. 32(5), 516-525.
Roll, R. (1984). Orange juice and weather. American Economic Review, Vol.
74, pp. 861-880.
177
Roques, F., Hiroux, C., Saguan, M. (2009). Optimal wind power deployment in
Europe
–
A
portfolio
approach.
Energy
Policy,
doi:10.1016/j.enpol.2009.07.048.
Rosenthal, L. and Young, C. (1990). The seemingly anomalous price behavior
of Royal Dutch/Shell and Unilever N.V./PLC. Journal of Financial Economics,
Vol. 26, 123-141.
Rowlands, I.H. (2007). The development of renewable electricity policy in the
Province of Ontario: the influence of ideas and timing. Review of Policy
Research, Vol. 24(3), 185-207.
Rowlands, I.H. (2005). Envisaging feed-in tariffs for solar photovoltaic
electricity: European lessons for Canada. Renewable and Sustainable Energy
Reviews, Vol. 9(1), 51-68.
Russo, M.V. (2003). The emergence of sustainable industries: building on
natural capital. Strategic Management Research Journal, Vol. 24, 317-331.
Russo M.V. and Fouts, P.A. (1997). A resource-based perspective on
corporate environmental performance and profitability. Academy of
Management Journal, Vol. 40(3): 534-559.
Sammer, K. (2007). Der Einfluss von Ökolabelling auf die Kaufentscheidung Evaluation der Schweizer Energieetikette mittels Discrete-ChoiceExperimenten, doctoral dissertation nr. 3404, University of St. Gallen,
Graduate School of Business, Economics, Law and Social Sciences.
Samuelson, P.A. (1965). Proof That Properly Anticipated Prices Fluctuate
Randomly. Industrial Management Review, Vol. 6(2), 41-49.
Saponar, C. (2010). Renewable Energy. Power Point presentation, IEA World
Energy Outlook workshop on “Outlook for renewable energy”, Paris, 26 April
2010.
Savage, L.J. (1954). The Foundations of Statistics, New York: Wiley.
Sawin, J.L. (2004). National Policy Instruments: Policy lessons for the
advancement and diffusion of renewable energy technologies around the
world. Thematic Background Paper for the International Conference on
Renewable Energies, Bonn, Germany, June 2004.
Sawtooth (2007). The ACA/Web v6.0 Technical Paper. Sawtooth Software
Technical Paper Series, Sawtooth Software Inc., Sequim, WA.
Sawtooth (2006). The ACA/Hierarchical Bayes v3.0 Technical Paper.
Sawtooth Software Technical Paper Series, Sawtooth Software Inc., Sequim,
WA.
Sawtooth (2002). CVA/HB Technical Paper. Sawtooth Software Technical
Paper Series, Sawtooth Software Inc., Sequim, WA.
Seale, O. (1999). The quality of qualitative research. London, Sage
Publications.
178
Sewell, M. (2007). Behavioural Finance. University of Cambridge, February
2007 (revised April 2010).
Sharpe, W.F. (1964). Capital Asset Prices: A Theory of Market Equilibrium
Under Conditions of Risk. Journal of Finance, Vol. 19(3), 425-442.
Shepherd, D.A. and Zacharakis, A. (1999). Conjoint analysis: A new
methodological approach for researching the decision policies of venture
capitalists. Venture Capital, Vol. 1, 197-217.
Shiller, R. (1981). Do stock prices move too much to be justified by
subsequent changes in dividends? American Economic Review, Vol. 71, 421436.
Shleifer, A. (2000). Inefficient Markets – An introduction to Behavioural
Finance. Clarendon Lectures in Economics, Oxford University Press, Norfolk.
Shleifer, A. and Vishny, R. (1997). The limits of arbitrage, Journal of Finance,
Vol. 52, 35-55.
Sijm, J.P.M. (2002). The performance of feed-in tariffs to promote renewable
electricity in European countries, ECN-C-02-083, November 2002.
Simon, H., (1957). A Behavioral Model of Rational Choice, in: Models of Man,
Social and Rational: Mathematical Essays on Rational Human Behavior in a
Social Setting. New York: Wiley.
Sitkin, S.B. and Pablo, A.L. (1992). Reconceptualizing the determinants of risk
behavior, The Academy of Management Review, Vol. 17(1), 9-38.
Snow, C.C. and Thomas, J.B. (1994). Field research methods in Strategic
Management: Contributions to theory building and testing. Journal of
Management Studies, Vol. 31(4), 457-480.
Sonntag-O’Brien, V. and Usher, E. (2004). Mobilising Finance for Renewable
Energies, Thematic Background Paper for the International Conference on
Renewable Energies, Bonn, Germany, June 2004.
Sovacool, B.K. (2009). The importance of comprehensiveness in renewable
electricity and energy-efficiency policy. Energy Policy, Vol. 37(2009), 15291541.
Staw, B.M., Sandelands, L.E. and Dutton, J.E. (1981). Threat-rigidity effects in
organizational behavior: A multi-level analysis. Administrative Science
Quarterly, Vol. 26, 501-524.
Stern, N. et al. (2006). Stern Review on the Economics of Climate Change,
Cambridge University Press.
Sung Choi, Y. and Whi Lee, B. (1995). Analysis of policy alternatives on the
public acceptance of nuclear power plant in Korea. Journal of Nuclear Science
and Technology, Vol. 32(3), 218-225.
179
Tanaka, N. (2010). The role of Solar Energy in Decarbonising the Electricity
Sector. Presentation held at the Mediterranean Solar Plan Conference,
Valencia, 11 May 2010.
Tanaka, N. (2008). Transition to Renewables. Presentation held at the
Ministerial Level Plenary Session of the Washington International Renewable
Energy
Conference (WIREC 2008). Washington, March 2008,
(http://www.iea.org/Textbase/speech/speechesresults.asp?SPEAKER_ID=63
6).
Teppo, T. (2006). Financing Clean Energy Market Creation – Clean Energy
Ventures, Venture Capitalists and Other Investors, doctoral dissertation,
series 2006/1, Helsinki University of Technology.
Thaler, R.H. (1999). The End of Behavioral Finance. Financial Analysts
Journal, November/December.
Thaler, R.H. and Johnson, E.J. (1990). Gambling with the house money and
trying to break even: The effects of prior outcomes on risky choice.
Management Science, Vol.36, 643-660.
Tsang, E.P.K. (2008). Computational intelligence determines effective
rationality. International Journal on Automation and Control, Vol. 5(1), 63-66.
Tversky, A. (1967). A General Theory of Polynomial Conjoint Measurement.
Journal of Mathematical Psychology, Vol.4, 1-20.
Tversky, A., and Kahneman, D. (1981). The Framing of Decisions and the
Psychology of Choice, Science. New Series, Vol. 211(4481), 453-458.
Tversky, A. and Kahnemann, D. (1974). Judgement under Uncertainty:
Heuristics and Biases, Science. New Series, Vol. 185(4157), 1124-1131.
UNEP FI (2009). 2009 Investor statement on the urgent need for a global
agreement on climate change (retrieved from http://www.unepfi.org).
UNEP and Bloomberg NEF – New Energy Finance (2010). Global Trends in
Sustainable Energy Investment 2010, United Nations Environment
Programme and New Energy Finance.
UNEP and Bloomberg NEF – New Energy Finance (2009). Global Trends in
Sustainable Energy Investment 2009, United Nations Environment
Programme and New Energy Finance.
UNFCCC (2007). Investment and Financial Flows to Address Climate
Change, Bonn.
Usher, E. (2008a). Creating the Climate for Change – Building Capacities to
Mobilise Investments, background document for the Tenth Special Session of
the Governing Council of UNEP / Global Ministerial Environment Forum, The
Principality of Monaco 20-22 February 2008.
Usher, E. (2008b). Global Investment in the Renewable Energy Sector. In:
IPCC – Working Group III, IPCC Scoping Meeting on Renewable Energy
180
Sources, Proceedings of the meeting held in Lübeck, Germany, 20-25
January 2008.
Van Zandt, T. (1999). Real-time decentralized information processing as a
model of organizations with boundedly rational agent’. Review of Economic
Studies, Vol. 66 (228).
Verhaegen, K., Meeus, L., Belmans, R. (2009). Towards an international
tradable green certificate system – The challenging example of Belgium.
Renewable and Sustainable Energy Reviews, Vol. 13(1), 208–215.
von Neumann, J. and Morgenstern, O. (1944). The Theory of Games and
Economic Behavior, Princeton, N.J.: Princeton University Press.
Voogt, M., Boots, M.G., Schaeffer, G.J., Martens, J.W. (2000). Renewable
electricity in a liberalised market: the concept of green certificates. Energy and
Environment, Vol. 11(1).
Vranas, P.B.M. (2000). Gigerenzer’s normative critique of Kahneman and
Tversky. Cognition, Vol. 76, 179-193.
Wang, Y. (2006). Renewable electricity in Sweden: an analysis of policy and
regulations. Energy Policy, Vol. 34, 1209-1220.
Weber, O., Köllner, T., Habegger, D., Steffensen, H. and Ohnemus, P. (2005).
The relation between sustainability performance and financial performance of
firms, GOE report No.5-2005, GOE, Zurich (www.goe.ch) and ASSET4
(www.asset4.com).
WEF – World Economic Forum (2010). Green Investing 2010 – Policy
Mechanisms to Bridge the Financing Gap, Geneva and New York, January
2010.
Weller, J. (2008). Best practices and lessons learned in drafting Renewable
Energy law in EU. Power point presentation, US AID Georgia, 11-13 July
2008.
Whitmore, G.A. and Cavadias, G.S. (1974). Experimental determination of
community preferences for water quality-cost alternatives. Decision Sciences,
Vol. 5, 614-631.
Wind, Y. and Spitz, L.K. (1976). An Analytical Approach to Marketing Decision
in Health Care Organizations. Operations Research, Vol. 24, 973-990.
Wiseman, R.M. and Gomez-Mejia, L.R. (1998). A Behavioral Agency Model of
Managerial Risk Taking, The Academy of Management Review, Vol. 23(1),
133-153.
Wiser, R. and Pickle, S. (1997). Financing investments in renewable energy:
the role of policy design and restructuring. University of California, Ernest
Orlando Lawrence Berkeley National Laboratory, Environmental Energy
Technologies Division, LBNL-39826, UC-1321, March 1997.
181
Wittink, D.R. and Cattin, P. (1989). Commercial Use of Conjoint Analysis: An
Update. Journal of Marketing, Vol. 53(3), 91-96.
Wright, R.W. (2002). Private capital and private markets. In: Sayigh, A.A.M.
(Ed), World Renewable Energy Congress VII (WREC 2002), Koln, July 2002.
Elsevier, Amsterdam.
Wüstenhagen, R., Wolsink, M. and Bürer, M.J. (2007). Social acceptance of
renewable energy innovation: An introduction to the concept. Energy Policy,
Vol. 35(5), 2683-2691.
Wüstenhagen, R., Wuebker, R., Bürer, M.J., Goddard, D. (2009). Financing
fuel cell market development: Exploring the role of expectation dynamics in
venture capital investment. In: Pogutz, S., Russo, A., Migliavacca, P. (Eds.):
Innovation, Markets, and Sustainable Energy: The Challenge of Hydrogen and
Fuel Cells, Edward Elgar: Cheltenham (UK) and Lyme (US), 118-137.
Yin, R.K. (2003). Case study research: design and methods. Third Edition,
Thousand Oaks, California: Sage Publications.
Young, F.W. (1969). Polynomial Conjoint Analysis of Similarities: Definitions
for a Special Algorithm. Research Paper No. 76, Psychometric Laboratory,
University of North Carolina.
Zacharakis, A.L. and Shepherd, D. (2001). The nature of information and
overconfidence on venture capitalists’ decision making. Journal of Business
Venturing, Vol. 16, 311-332.
182
Annex 1: The questionnaire
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
Annex 2: List of companies contacted
France
ACE Management
Arcis Group
Auriga Partners
Bryan, Garnier and
Co.
CDC Innovation
Crédit Suisse France
Debevoise and
Plimpton
AGF PE
A Plus Finance
Axa
Calyon
CPR Private Equity
CVC Capital Partners
France
Demeter
Innoven
La Compagnie du
Vent
Emertec Energie
Environnement
Groupama Asset
Management
Iris Capital
Olympia Capital
Management
Rothschild
Sinopia (HSBC)
Sofinnova
Truffle
Xange
TechFund
Ventech
21 Centrale Partners
ECF
Fortis PE France
Alven Capital
Atlas Venture
BNP Paribas
Capital Fund
Management
Crédit Agricole
Dalkia
Dexia
Eraam
HDF
Ixis
Partech
Société Generale
Asset Management
Theolia
Viveris Management
3i Gestion SA
Germany
Advent International
Bonventure GmbH
B.A.U.M. Group
Capiton AG
Deutsche Bank
Earlybird
EUtech Energie &
Management
Extorel
First Solar GmbH
HassoPlattner
Ventures
KfW
Neuhaus Partners
Seed Capital
Brandenburg GmbH
Target Partners
High-Tech
Gründerfonds
Mountain Cleantech
NRW Bank
Wellington Partners
West LB
Greenstream
HVB/Unicredit
Munich RE
PNE Wind
Star Ventures
Ventizz Capital
Partners
White&Case
Baytech
Cipio Partners
eCAPITAL
entrepreneurial
Partners AG
Sj Berwin
Technostart GmbH
3i Deutschland Private Equity
205
Italy
ABN Amro Capital
Advanced Capital
Alto Partners
Accord Management
Advent International
Argan Capital
Advisors
Argy Venture Capital
AVM Private Venture
BC Partners
BCC Private Equity
Centrobanca SpA
CI Partners
Cinven
CIR Ventures
DGPA SGR
Efibanca
Fidi Toscana
Finlombarda
Equiter
Fidia SGR
Fises
Gruppo Banca
Leonardo
ICQ Holding
Interbanca SpA
Friulia
IA Partners Srl
Innogest SGR
IP Investimenti e
Partecipazioni
MPS
Palladio Finanziaria
Quantica
Hat Holding
IMI Investimenti
Intesa San Paolo
Orizzonte SGR
Permira Associati
PAI Partners
Pino Partecipazioni
S+R Investimenti e
Gestioni
ReFeel
The Carlyle Group
TT Venture
Yarpa
AXA Private Equity
Italy
BNP Paribas Asset
Management
Cimino&Associati
Private Equity
Clessidra SGR
Doughty Hanson &
Co.
F2i
Filas
Fondamenta SGR
Mandarin Advisory
Sofipa
Vertis
Argos Soditic Italia
L Capital Advisory
San Paolo IMI
Investimenti per lo
Sviluppo SGR
Solar Ventures
San Paolo IMI Fondi
Chiusi
Actelios Falck
Alcedo SGR
Vestar Capital
Partners Italia
21 Investimenti
Sigefi Italia Private
Equity
Strategia Italia SGR
UniCredit corporate
banking
Wise SGR
3i SGR
206
Spain
Activos y Gestión
Accionaria
Adara
AIG
Altamar Capital
Axon Capital e
Inversiones
Banking Capital
Riesgo
Bridgepoint Capital
Caja Madrid
Cidem
Clave Mayor SA
Corsabe Corporación
Sant Bernat
Demeter Partners
Eland Private Equity
Limited
Diana Capital
GED Iberian PE
Green Alliance
Grupo Santander
Sadim Inversiones SA
Sepides
SPPE
Torsa Capital
Valcapital Gestión
YSIOS Capital
Partners
Varde Europe
Aeris Capital
Alpha Associates AG
Axon Partners
AIG Investments
Aravis
Baker & McKenzie
BC Partners
Bridgelink AG
CTI Invest
EPS Value Plus
Invision
Mountain Partners
New Value
SAM Services AG
Syngenta Ventures
Emerald Technology
Good Energies
LGT Capital
NanoDimensions
Partners Group
Silverdale
Venture Partners
Baring Private Equity
Partners España
Capital Alianza Private
Equity Investment
Corphin Capital
Asesores
Cygnus AM
EBN Capital
Going Investment
Gestion
Qualitas Equity
Partners
SES Iberua Private
Equity
Valanza
XesGalicia
3i Europe plc
Switzerland
Adveq Management
AIG Private Equity
Argos Soditic
Barclays Private
Equity
BV Group
Endeavour Vision
Index Venture
Luserve
New Energies
Quadrivium
Swiss RE
Zurmont Madison
Management
3D Capital
207
United Kingdom
Advent Venture
Partners
AON
Atlas Venture
Black River Clean
Energy
Brit Syndicater Lloyds
CC Capital
Citigroup
Conduit Ventures
Doughty Hanson
E-Synergy
Foursome
GE Capital
Headway Capital
Partners
Hunton & Williams
London Asia Capital
Aloe Private Equity
Amadeus Capital
ARC InterCapital
Balderton Capital
Argo Capital
Barclay
BlackRock
BP
Camco
Cheyne Capital
Cleantech Venture
Network
Consensus Business
Group
Englefield Capital
Eversheds LLP
Frontiers Capital
Partners
Goldman Sachs
Carbon Trust
Cisco Systems Europe
Climate Investment
Partners
HG Capital
HSBC
IP Group plc
Macquarie
ISIS Equity Partners
Man Group
New Star Asset
Management
Octopus Investments
Pond Ventures
SEP
Merrill Lynch
Morgan Stanley
Nomura
Oxara Energy
Royal Bank of Canada
North Star
Platina
RREEP Private Equity
Sindicatum Carbon
Capital
TLCom
WestLB AG
Silverdale
Standard Charter
Virgin Green Fund
Zouk Ventures
Other countries
Attica Ventures
Clean World Capital
Invest Equity
Capital-E Partners
EIB
Global Cleantech
Capital
Piraeus Bank
Sitra
Stonefund
Fortis
DFJ Esprit
Entrepreneur’s Fund
Fidelity Venture
Gartmore Investment
Management
Good Energies
Standard Bank
Unicredit Group
WHEB Ventures
Capricorn
Energi Invest Fyn A/S
GloCal Systems
Quest Management
Strategus
Management
208
Annex 4: Curriculum vitae
Surname / First name
Address
Telephone
E-mail
Nationality
Date of birth
Gender
MENICHETTI Emanuela (Mrs)
3 Cité de Phalsbourg 75011 Paris, France
+33 143 56 22 94
[email protected]
Italian
26.07.1973
Female
Emanuela Menichetti is PhD candidate at the University of St. Gallen,
Institute for Economics and the Environment, under the supervision of Prof. Dr.
Rolf Wüstenhagen and the co-supervision of Prof Dr. Rolf Peter Sieferle. Her
doctoral dissertation focuses on policy risk perceptions and investors’ decision
making behaviour in the renewable energy finance sector.
Mrs. Menichetti has over 10 years of working experience in the field of
renewable energy and sustainability. Since September 2008 she has been
serving as Senior Energy Analyst at the Observatoire Méditerranéen de
l’Energie, an association of some 30 leading energy companies operating in
the Mediterranean basin, located in France. Prior to this position, she worked
as Associate Programme Officer at the Energy Branch of the United Nations
Environment Programme (UNEP) Division of Technology, Industry and
Economics (from December 2006 to August 2008), and as a researcher and
consultant at Ambiente Italia, a leading environmental policy and analysis
institute in Italy (from March 1999 to November 2006). Mrs. Menichetti is
author or co-author of over 20 reports and about 15 publications.
Mrs. Menichetti enjoys a multidisciplinary background. After graduating
summa cum laude in Economics and Business Administration with a thesis on
Economics and Policy for the Environment in 1998, she completed an
additional Master in Environmental Economics (final mark: 30/30) in 1999.
An Italian citizen, Mrs. Menichetti speaks also English, French and Spanish.
Since 2007 she has been studying Arabic at UNESCO.
Mrs. Menichetti is a certified C++ programmer and is able to work with various
softwares for environmental analysis. She has a good command of the
Sawtooth software for computer-administered surveys and conjoint analysis,
and a basic knowledge of SAS and SPSS for statistical analysis.
209
Statutory Declaration
Hereby I declare
- that I wrote this dissertation without any illicit assistance and without using
any other aids than stated and that this dissertation was neither presented in
equal nor in similar form at any other university;
- that I cited all references that were used respecting current academic rules.
Place and date of issue: Paris, 1 June 2010
Signature:
210

Renewable Energy Policy Risk and Investor Behaviour

Transcrição

Documentos relacionados

Renewables Club - Communiqué - The International Renewable

5th Renewable Energy Conference, Hamburg

5. Energia

who is behind ere international?

PRESS RELEASE For immediate release: 28 October

to - Enersolar + Brasil

to - Enersolar + Brasil

Renewable energy—traditional biomass vs. modern biomass

www .zfu.ch

ASEAN Sustainable Energy Week 2016